List papers
| Seq | # | Title | Abstract | Keywords | DOIX | Authors with affiliation and country |
|---|---|---|---|---|---|---|
| Seq | # | Title | Abstract | DOI | Keywords | Authors with affiliation and country |
| 1 | 1570875973 | A Secure Cloud Framework for Big Data Analytics Using a Distributed Model | The paper aims to provide a framework for the secure processing and analysis of big data. It employs a double layer of security that is based on Elliptical Curve Cryptography (ECC) and Fully Homomorphic Encryption (FHE). FHE can be utilized as a capable public-key encryption that permits users to do computations on encrypted data. Additionally, a clustering method has been used to partition big data into smaller data sizes that are distributed among many virtual CPUs. In the defined distributed model, many VMs process different partitions of data parallelly and simultaneously to reduce the processing time of data. KMeans clustering algorithm is used in three datasets as an instance of data analytics to test the suggested framework. Furthermore, the produced results are compared with a centralized-based model to assess the productivity and efficiency of the distributed model. Besides, the \textit{principal component analysis} (PCA) is applied to the used clustering algorithm to diminish the required clustering time needed by the distributed model. The results indicate that the clustering time can be reduced up to 91\%, and even with 18\% more reduction in the execution time using the distributed model. The recommended solution can improve the effectiveness of big data analytics whereas guaranteeing the security of such data. | http://dx.doi.org/10.12785/ijcds/150101 | Cloud security; big data analytics; hybrid encryption; KMeans clustering; distributed computation | Zainab Salman (University of Bahrain & IT College, Bahrain); Alauddin Yousif Al-Omary (University of Bahrain, Bahrain); Mustafa Hammad (Mutah University, Jordan) |
| 2 | 1570899198 | Ontological Concepts and Logical Listeners for Dysarthric Speech Understanding | Depending on the social features of the speaker and the social setting in which they are speaking, the relationship between meaning and context might alter. In order to interpret the meaning from dysarthric speech, this paper proposes a theoretical framework for employing speech-event representations, also known as situational projections. The multi-layered approach has been broken down into four main components: a few-shot learner that builds up speaker familiarity; a situational projection component that marshals natural sentences and the built-up familiarity markers into a vector triple; a contextualizer that builds up ontological concepts of the input triple; and finally, a transducer that assumes the function of a logical listener. | http://dx.doi.org/10.12785/ijcds/150102 | Dysarthria; Speech meaning; Speech context; Familiarity; Situational Projection; Dialogue Acts | Benard Alaka and Bernard Shibwabo Kasamani (Strathmore University, Kenya) |
| 3 | 1570900957 | Global Stock Price Forecasting during a Period of Market Stress using LightGBM | Accurately predicting future stock prices is crucial for investors, particularly during market stress, enabling informed decisions to mitigate losses and reduce financial exposure. While machine learning techniques have shown promise in this field, most studies have focused on local models. Global forecasting models, which are trained on a variety of time series, have demonstrated encouraging outcomes in outperforming local models. This article presents a global forecasting approach utilizing LightGBM to predict stock prices in the Moroccan market during highly volatile period. The study includes a sector-wise analysis, with the most volatile sector being evaluated, and correlated stocks from other sectors were considered to enhance the data. Compared to other deep learning models that use a local approach, our findings demonstrate the effectiveness of utilizing a global forecasting with optimized LightGBM for predicting stock prices. Additionally, incorporating data from multiple stocks improves accuracy of stock price prediction. | http://dx.doi.org/10.12785/ijcds/150103 | Stock market; Market stress; LightGBM; Global forecasting | Omaima Guennioui (Université Mohammed V de Rabat, Morocco); Dalila Chiadmi (Mohammadia School of Engineering & Mohammaed V Agdal University, Morocco); Mustapha Amghar (Ecole Mohammadia des Ingenieurs, Morocco) |
| 4 | 1570901656 | Channel Modeling Techniques for Multiband Vehicle to Everything Communications: Challenges and Opportunities | Vehicle to everything (V2X), one of the most popular fifth generation (5G) application areas, demands a high data rate, extremely low latency, and high reliability. Investigations that enable exploiting the synergistic potential of the millimeter wave (mmWave) and sub-6 GHz frequencies play a prominent role in meeting these goals. The wireless channels through which communication signals propagate primarily determines overall system performance. Hence, a detailed study of the propagation channel behavior of these bands becomes an essential step to compare goals set by standards bodies with that of practically achievable performances. This evaluation yields a clue to select appropriate complementing technologies that can minimize the gaps. Currently, multidimensional research works are being carried out to overcome constraints at wireless channels and redeem the numerous potential of joint sub-6 GHz/mmWave bands. In this survey article, findings from existing works on V2X communication channel modeling approaches are summarized by focusing on challenges unaddressed and opportunities available. This paves ways to assess achieved targets and remaining ones with respect to the requirements of 5G multiband vehicular communications. Besides concisely organizing existing works, the role of the key technologies in achieving multiband-based V2X targets and procedures to schedule users when they dwell in a region where both bands co-exist are also highlighted. From an overall study, challenges existing till now and opportunities in hand to realize joint sub-6 GHz/mmWave frequency bands-based services for the near future luxurious vehicles are pointed out. Therefore, the autonomous driving potential of joint sub-6 GHz/mmWave can be realized when the channel dynamics are adequately modeled hence further studies considering the realistic propagation environment are required. | http://dx.doi.org/10.12785/ijcds/150104 | Channel Model; Propagation loss; Millimeter Wave; Multiband; Vehicular Communication | Degefe S Hisabo (AAiT, Ethiopia); Thomas Otieno Olwal (Tshwane University of Technology, South Africa); Murad Ridwan Hassen (Addis Ababa Institute of Technology, Addis Ababa University, Ethiopia) |
| 5 | 1570905771 | A Fusion Architecture of BERT and RoBERTa for Enhanced Performance of Sentiment Analysis of Social Media Platforms | Natural language processing's subfield of sentiment analysis involves locating and categorizing the feelings, viewpoints, and attitudes expressed in text. Because it enables us to understand public opinion on a variety of topics, sentiment analysis has grown in importance as social media platforms become more widely used. In this research paper, we used two cutting-edge deep learning models, BERT and RoBERTa, and their fusion of both architectures to perform sentiment analysis on a dataset of tweets related to the COVID-19 pandemic. To eliminate noise and unrelated data, the dataset underwent pre-processing and cleaning. Then, using the dataset, we trained the BERT and RoBERTa models and assessed their performance. For sentiment analysis, both models achieved high F1-scores, recall, and accuracy for all three sentiment classes (negative, neutral, and positive). While there were some differences in how well these models performed across these metrics, both models did well and classified the sentiment of tweets in the dataset with high accuracy. Our study's findings show how well BERT and RoBERTa perform sentiment analysis on tweets about the COVID-19 pandemic. Our study also emphasizes how crucial it is to clean up and pre-process the dataset to get rid of extraneous data and noise that can harm the models' performance. The effectiveness of these models on datasets from other domains and topics can be investigated in further research. Future studies should also look into the models' interpretability and comprehend the features and patterns crucial to sentiment analysis. | http://dx.doi.org/10.12785/ijcds/150105 | Unsupervised learning; Sentiment Analysis; BERT; RoBERTa; Deep learning; performance parameters | Pranay BV Kumar (Kakatiya University, India); Manchala Sadanandam (Kakatiya University Warangal India, India) |
| 6 | 1570910069 | A Stable Method for Brain Tumor Prediction in Magnetic Resonance Images using Finetuned XceptionNet | Brain tumors can be a life-threatening condition, and early detection is crucial for effective treatment. Magnetic resonance imaging (MRI) is a valuable appliance for identifying the tumor's location, but manual detection is a time-engrossing and flaws-prone process. To overcome these challenges, computer-assisted approaches have been developed, and deep learning (DL) archetypes are now being pre-owned in medical imaging to discover brain tumors manoeuvre MRI carbon copies. In this, we propose a deep convolutional neural network (CNN) Xception net model for the efficient classification and detection of brain tumor images. The Xception net is a powerful CNN model that has shown promising results in various systems perceiving exercise, in conjunction with medical illustration scrutiny. We fine-tuned the Xception net model using a dataset of Magnetic Resonance Imaging (MRI) images of the brain, which were pre-processed and labeled by medical experts. To reckon the performance of our prototype, we counselled dossier using a variety of interpretation criterion, including accuracy, precision, recall, and F1 score. Our causatums view that the urged model achieved high accuracy in classifying brain tumor images. The archetype's strength to accurately and efficiently classify and detect brain tumors using MRI images can significantly improve patient outcomes by enabling early detection and treatment. Overall, our study demonstrates the persuasiveness of using the Xception net flawless for brain tumor ferreting out and alloting using MRI images. The proposed model has the potential to revolutionize the department of salutary exemplify and improve patient outcomes for brain tumor treatment. | http://dx.doi.org/10.12785/ijcds/150106 | Brain Tumor; Deep Convolution Neural Networks; Magnetic Resonance Imaging; XceptionNet | Shanmuga Sundari M, Divya Yeluri, Durga Kbks, Vidyullatha Sukhavasi, Dyva Sugnana Rao M and Sudha Rani M (BVRIT HYDERABAD College of Engineering for Women, India) |
| 7 | 1570912257 | Introduction of LSSVR for the Prediction of the Yellowness Index | Principal component regression (PCR), partial least square regression (PLSR), and least square support vector regression (LSSVR) have been widely applied as predictive models in various applications. However, studies employing regression models to estimate the yellowness index (YI) are scarce in the literature. This study, therefore, focuses on developing non-destructive YI measurements using regression models. The collected RGB calculated XYZ and obtained CIE LAB values were set as the input variables. Meanwhile, the YI value was denoted as the output variable. Results indicated that the LSSVR model outperforms PCR and PLSR models in predicting YI in which the root means square errors of LSSVR for training and testing datasets were found to be 261,406% to 294,218% and 725% to 772% lower than PLSR and PCR, respectively. LSSVR is also attributed to higher coefficients of determination (R2) for training and testing datasets compared to PLSR and PCR, whose R2 values are very close to 1. Nonetheless, the computational times of training and testing datasets for LSSVR are much longer than that of PLSR and PCR. Consequently, a highly accurate LSSVR model-based YI sensor shows promising applications particularly if the computational load can be further minimized. | http://dx.doi.org/10.12785/ijcds/150107 | Least square support vector regression; Yellow Color; Least square regression model; Partial least square regression; Principal component regression | Wan Sieng Yeo (Curtin University Malaysia, Malaysia); Agus Saptoro (Curtin University, Malaysia) |
| 8 | 1570912877 | Techno-Economic and Environmental Benefit Analysis of PV and D-FACTS Enriched Modern Electric Distribution Grid with Practical Loads | In the current era, expansion in the power system has led to an increase in the Distributed Generation (DG) planning to accomplish the increase in demand. Hence because of the opted DG's inappropriate planning, the system gets unbalanced with an increase in losses, voltage profile imbalance along with reliability issues. Here, the modernized IEEE 69-bus network is considered for appropriate DG's planning. In this study, one important factor, which is the Line Fault Current Level (LFCLevel) along with Real & Reactive Losses (RLoss & QLoss), Voltage Deviation Profile (VDP), and Network Reliability (RN), is considered as an index in the objective function, which makes the proposed Multi-Objective Function (MOF) a novel MOF (nMOF). The cost-economic and Pollutant Gas Emissions (PGE) parameters are also examined in accordance with this nMOF optimized outcomes. In the proposed study, the DG's (PV and D-FACTS) are planned optimally with reconfiguration incorporating practical loads (industrial, commercial and residential) using an Adaptive Particle-Swarm Optimization approach (APSO). Then the techno-economic and environmental-benefit analysis of the results reveals that the losses (real and reactive), cost, and pollutant gas emissions are reduced with an improved balanced profile of voltage, LFCLevel, and RN. | http://dx.doi.org/10.12785/ijcds/150108 | Photovoltaic (PV); Distributed-Flexible AC Transmission System (D-FACTS); Tie-Switches (TS); novel Multi-Objective-Function (nMOF); Pollutant Gas Emissions (PGE); Adaptive Particle Swarm Optimization (APSO) | Bikash Kumar Saw (National Institute of Technology Durgapur, West Bengal, India); Aashish Kumar Bohre (NIT Durgapur, India) |
| 9 | 1570920947 | Decision Tree Induction Using Evolutionary Algorithms: A Survey | An evolutionary methods for an induction-based decision trees made a wide step development in machine learning field. In this context, the majority of researches recently concentrates on techniques that use developing decision trees as an alternative to the traditional heuristic top-down divide-and-conquer strategy. Evolutionary algorithms play an important role in improving decision tree classifier parts. The main contributions of our article are two folds, first it provides a survey of evolutionary algorithms and decision trees. Second, it reviews a taxonomy that encompasses techniques mentioned above as a backbone in creating enhanced decision trees, and those that evolve the construction components of decision trees. The article covers researches in the period 2011-2023. Finally, a detailed scenarios and results had been analyzed, and compared to spot the lights on those strength points and defects with respect to processing time, accuracy, and required space | http://dx.doi.org/10.12785/ijcds/150109 | Classification; Decision tree; Evolutionary algorithms; Genetic algorithms; Machine learning | Maryam Hussein Bahar (University of Babylon, Iraq); Hadeel Noori Saad (University of Kufa, Iraq) |
| 10 | 1570927370 | GRNRI: Gene Regulatory Network Inference using Unsupervised Graph Neural Network | The task of reconstructing Gene Regulatory Networks (GRNs) from gene expression data is a complex and challenging problem in the field of systems biology. In our paper, we present GRNRI, an innovative unsupervised model that is capable of inferring GRNs from single-cell RNA sequencing data. Specifically, we have developed a modified version of the Neural Relational Inference (NRI) approach that can explicitly model the regulatory relationships between genes. Our model is a variational auto-encoder where the latent code represents the GRN and the reconstruction uses graph neural networks. Our results demonstrate that GRNRI achieves comparable or superior performance on most benchmark datasets when compared to state-of-the-art methods. | http://dx.doi.org/10.12785/ijcds/150110 | Gene Regulatory Network; Single Cell RNA sequencing; Graph Neural Network; Unsupervised Learning; Bioinformatics | Fatema Abdulhameed and Hazem Abbas (Ain Shams University, Egypt); Cherif R. Salama (Ain Shams University, Egypt and the American University in Cairo, Egypt) |
| 11 | 1570937977 | Optimization of Population Document Services in Villages using Naive Bayes and k-NN Method | The Indonesian government strives to improve its registration and the issuance of the population documents program. However, several obstacles are faced, such as the complicated topography of the area and the distance from the village. Therefore, ball pick-up services are urgently needed. The government of Alor Regency, East Nusa Tenggara Province, is one of the regions that has implemented this program. However, not all villages can be served due to limited time and funds. Therefore, villages must be selected fairly so the program can run well. Machine learning, a classification technique using data mining concepts, is expected to overcome this problem. This research aims to identify the most effective method for classifying eligible villages. The experimental process includes preprocessing, model training using K-NN and NB, and performance evaluation. The results show that both methods provide good results, albeit with slightly different levels of accuracy. Comparative analysis shows that the K-NN method has a higher accuracy rate of 97.14% for k=1 and k=2 on the MMN-normalized dataset but has the lowest accuracy of 77.1% at k=11 and k=13 on the raw dataset. In comparison, the NB method has an accuracy of 94.29% but is stable on raw and normalized datasets. | http://dx.doi.org/10.12785/ijcds/150111 | Data mining; Machine learning; K-NN; Naive bayes; Comparative analysis; village selection | Imam Riadi (Department of Information System, Universitas Ahmad Dahlan, Yogyakarta 55164, Indonesia); Anton Yudhana (Departement of Electrical Engineering, Universitas Ahmad Dahlan, Yogyakarta 55164, Indonesia); M Rosyidi Djou (Master Program of Informatics, Universitas Ahmad Dahlan, Yogyakarta, 55164 Indonesia) |
| 12 | 1570940891 | Design of E-Portfolio to Enhance Career Development for TVET Graduates | Technical and Vocational Education and Training (TVET) is an education sector that prepares individuals with the skills required by the industry to improve the individuals' employability and competitiveness as workforce at national and international levels. The use of information technology as education technology in TVET supports the integration of technological advancement with TVET components in order to assure the compatibility between the graduates' skills and the industrial needs. E-Portfolio as one fruit of the advancement of knowledge and technology offers a new solution to improve the absorption of TVET graduates as the workforce into the industry. The usability of E-Portfolio is determined by the design of E-Portfolio development that should describe the skills and competences of the graduates more precisely to the industry with an expectation to increase the absorption of the graduates into the employment world. E-Portfolio in this article is designed using website-based technology that is supported with four features, namely Portfolio Insight, Portfolios, Portfolio Journey, and My Library. Navigation support system is included in the design of E-Portfolio and becoming innovative in portfolio development stages. The navigation support system takes a form as the textual guidance in the media users choose to exhibit their learning evidence. | http://dx.doi.org/10.12785/ijcds/150112 | E-Portfolio; career development platform; web application; innovation technology; system development; education technology | Shofiyah Al Idrus (State University of Malang & Fakultas Teknik, Indonesia); Rafiza Abdul Razak (University Of Malaya, Malaysia); Norharyanti Mohsin (Universiti of Malaya, Malaysia) |
| 13 | 1570942559 | Citrus Tree Nutrient Deficiency Classification: A Comparative Study of ANN and SVM Using Colour-Texture Features in Leaf Images | Nutrient deficiency in Citrus reticulata (mandarin orange) plants is a critical agricultural issue that has far-reaching implications. Such deficiencies not only compromise the overall health and vigor of the mandarin orange trees but also render them more susceptible to diseases and pest infestations. Addressing this concern is of principal importance to ensure sustainable citrus cultivation and enhance crop yield. In this research, we propose a comprehensive approach to tackle the pervasive problem of nutrient deficiencies in Citrus Reticulata var. Fremont leaves. Our approach leverages cutting-edge techniques in image processing and machine learning to provide accurate and efficient detection of these deficiencies. The image data is divided into four classes: N Deficiency, P Deficiency, K Deficiency and Normal. The file sizes are compressed using a lossless compression method, resulting in an average file size reduction of 96.99%. The second stage involves applying segmentation processes to the images using the Canny and Sauvola methods.Third stage involves extracting colour and texture features from the images. The feature values will be used for the classification process in the next stage. The segmentation process employs two methods, namely Canny and Sauvola, which effectively separate the leaves from the background. The detection process is evident in the feature extraction phase, which utilizes two features: colour and texture.The fourth stage involves the classification process based on the segmentation results methods, performed separately using Artificial Neural Network (ANN) and Support Vector Machine (SVM) methods. These process results in four datasets: Canny-ANN, Canny-SVM, Sauvola-ANN, and Sauvola-SVM. The highest accuracy is achieved by the Sauvola-ANN method, with a value of 93.75%. | http://dx.doi.org/10.12785/ijcds/150113 | ANN; citrus leaves; Canny; classification; Sauvola; SVM | Lia Kamelia (UIN Sunan Gunung Djati Bandung, Indonesia); Titik Abdul Rahman (Asia e University, Malaysia); Rin Rin Nurmalasari (UIN Sunan Gunung Djati Bandung, Indonesia); Kiki Kusyaeri Hamdani (National Research and Innovation Agency, Indonesia) |
| 14 | 1570949776 | AI-Empowered Flying Ad-Hoc Networks for Dynamic Connectivity | Flying Ad-Hoc Networks (FANETs) represent a dynamic and adaptable communication paradigm that leverages aerial nodes for establishing agile and resilient networks. The integration of Artificial Intelligence (AI) techniques into FANETs presents a revolutionary approach to addressing the challenges posed by dynamic connectivity in airborne scenarios. This paper provides a comprehensive exploration of AI-Empowered FANETs, highlighting their significance in overcoming mobility constraints, optimizing resource usage, enhancing security, and enabling novel applications. We delve into the transformative impact of AI on FANET capabilities, showcasing how AI-driven path planning, adaptive routing, and communication protocols revolutionize data exchange. Additionally, we explore AI's role in optimizing energy usage, enhancing security measures, and bolstering decision-making processes within FANETs. Through real-world applications and future directions, we illustrate the boundless potential of AI-Empowered FANETs in disaster response, environmental monitoring, surveillance, and beyond. By synthesizing the advancements and challenges, this paper underscores the profound impact of AI on FANETs, positioning them as a cornerstone of intelligent communication in dynamic airborne environments. | http://dx.doi.org/10.12785/ijcds/150114 | FANETs, Artificial Intelligence, AI techniques, Dynamic connectivity, Resource management, Energy optimization, Security enhancement | Belghachi Mohammed (Bechar University, Algeria) |
| 15 | 1570878114 | Prescriptive Analytical Models for Dynamic IoT Data Streams: A Review | The application of data analysis tools and procedures to percieve value from vast volume of data created by connected IoT devices is known as IoT data analytics. Many IoT applications are made up of disparate streams that must be merged and turned into complete, consistent, current, and accurate data for business reporting and analysis. While predictive analytics on IoT dealing with the prediction invloved with the setting of IoT appliances, the next stage of IoT data analytics maturity involves deriving actionable insights from predictions made in previous stages. The earlier descriptive, diagnostic, and predictive IoT analytics blog segments took the process one step closer to optimal decision-making, prescriptive analytics takes it the rest of the way. Prescriptive analytics is an emerging approach in the data analytics community that requires much attention to employ it in various fields of technology. It is a promising approach in IoT data analytics that focuses on discovering the distinguished action necessary for a particular circumstance, based on data. In this paper, an overview of IoT data analytics, survey of prescriptive analytical models, applications, issues, challenges and platforms for IoT analytics are discussed. | http://dx.doi.org/10.12785/ijcds/150115 | IoT data analytics; dynamic IoT data streams; predictive analytics; prescriptive analytics; optimal decision making | Shubha T V and Dilip Kumar Sm (Bangalore University, India); Venugopal Kuppanna Rajuk (University Visvesvaraya College of Engineering, India) |
| 16 | 1570880874 | A Detail Study of Sign Language Communication for Deaf-Mute to Normal Person | Even though sign language recognition systems are crucial, there isn't a well-structured literature study to help Deaf-Mute. 160 research articles were found and reviewed in this survey of sign recognition framework. Subsequently, around 100 research articles were chosen, analyzed, and categorized. Each of these articles was grouped according to several criteria, such as data-set, kind of sign language, data collection methods, processing, segmentation, learning strategy, hybrid or deep learning strategy, etc. Using a camera, the largest study on sign language recognition has been done on steady, separated, and one-handed signs, according to the literature review for this essay. Researchers are receiving increasing attention these days for creating cost-effective Sign Language Recognition. According to the review, the main reason why commercialized sign language hasn't yet been developed is that the current methods for identifying sign language either require expensive equipment or take too much computational time. The objective of the article is to study the techniques to identify the most effective methodology that has been employed by researchers. As a result, other researchers will be able to learn more about the techniques developed and possibly create future Sign Language Systems that are more effective in the field of computer vision. | http://dx.doi.org/10.12785/ijcds/150116 | Hand-Sign; Sign Language Recognition; Deep Learning; Computer Vision | Pareshbhai Mansangbhai Chauhan and Dineshkumar Bhagwandas Vaghela (Gujarat Technological University, India) |
| 17 | 1570890858 | Physical Key Generation Using Modified Discrete Wavelet Transforms for the Internet of Things | Security and privacy of data are important factors in the Internet of Things (IoT), which enables safe communication on emerging objects. The non-standardization of security protocols on the internet of things makes them vulnerable to intruders and compromises the process of communication. Conventional cryptography approaches face certain limitations in the generation of keys for authentication and authorization of communication systems. The limitations of the cryptography approach are explored in terms of computational overhead and hardware requirements. Singal processing is an alternative cryptography approach for key generation in physical-layer wireless communication systems. The wireless channel's reciprocity principle is used to generate keys. The important factor in a signal-based key generation is the process of quantization. In this manuscript, we propose modified discrete wavelet transform (DWT) methods for key generation in IoT. The modification of discrete wavelet transform methods encapsulates the process of common spatial patterning (CSP). The CSP is another signal-processing method for the formation of signal patterns. The proposed method reduces DWT discrepancies and yields a key generation technique that is efficient. Key generation processing used multi-bit quantization to quantify the production of the modified discrete wavelet transform. The proposed algorithm for key generation simulates different numbers of device nodes in indoor and outdoor scenarios. Use MATLAB tools to simulate the process and measure standard parameters such as AKL and KDR. According to the results, the suggested technique outperforms existing key generation algorithms by 3%. | http://dx.doi.org/10.12785/ijcds/150117 | IoTs; Physical Layer; Wireless Communication; DWT; CSP; Cryptography | Rakesh Sharma (National Institute of Technology & Kurukshetra, India); Poonam Jindal (NIT Kurukshetra, India); Brahmjit Singh (National Institute of Technology Kurukshetra, India) |
| 18 | 1570891417 | BushraDBR: An Automatic Approach to Retrieving Duplicate Bug Reports | A Bug Tracking System (BTS), such as Bugzilla, is generally utilized to track submitted Bug Reports (BRs) for a particular software system. Duplicate Bug Report (DBR) retrieval is the process of obtaining a DBR in the BTS. This process is important to avoid needless work from engineers on DBRs. To prevent wasting engineer resources, such as effort and time, on previously submitted (or duplicate) BRs, it is essential to find and retrieve DBRs as soon as they are submitted by software users. Thus, this paper proposes an automatic approach (so-called BushraDBR) that aims to assist an engineer (so-called a triager) to retrieve DBRs and stop the duplicates before they start. Where BushraDBR stands for Bushra Duplicate Bug Reports retrieval process. Therefore, when a new BR is sent to the Bug Repository (BRE), an engineer checks whether it is a duplicate of an existing BR in BRE or not via BushraDBR approach. If it is, the engineer marks it as DBR, and the BR is excluded from consideration for any additional work; otherwise, the BR is added to the BRE. BushraDBR approach relies on Textual Similarity (TS) between the newly submitted BR and the rest of the BRs in BRE to retrieve DBRs. BushraDBR exploits unstructured data from BRs to apply Information Retrieval (IR) methods in an efficient way. BushraDBR approach uses two techniques to retrieve DBRs: Latent Semantic Indexing (LSI) and Formal Concept Analysis (FCA). The originality of BushraDBR is to stop DBRs before they occur by comparing the newly reported BR with the rest of the BRs in the BTS, thus saving time and effort during the Software Maintenance (SM) process. BushraDBR also uniquely retrieves DBR through the use of LSI and FCA techniques. BushraDBR approach had been validated and evaluated on several publicly available data sets from Bugzilla. Experiments show the ability of BushraDBR approach to retrieve DBRs in an efficient and accurate manner. | http://dx.doi.org/10.12785/ijcds/150118 | Software engineering; Software maintenance; Duplicate bug reports; LSI; FCA | Ra'Fat Ahmad AL-msie'deen (Mutah University, Jordan) |
| 19 | 1570954882 | Convolutional Neural Network for Predicting Student Academic Performance in Intelligent Tutoring System | One of the most significant research areas in education and Artificial Intelligence (AI) is the earlier prediction of students' academic achievement. Limited studies have been conducted using Deep Learning (DL) in the student domain of Intelligent Tutoring System (ITS). Traditional Machine Learning (ML) techniques have been employed in many earlier publications to predict student performance. This paper investigates the effectiveness of DL algorithms for predicting student academic performance. Three different DL architectures based on the structure of Convolutional Neural Networks (CNN) are presented. Two public datasets are used. Furthermore, two feature selection techniques are utilized in this experiment: Principal Component Analysis (PCA) and Decision Trees (DTs). Also, we applied a resampling technique for the first dataset to address the issue of an imbalanced dataset. According to the experimental findings, the proposed CNN model's success in predicting student performance at early stages reached an accuracy of 94.36% using the first dataset and 84.83% using the second dataset. Comparing the proposed approach with the previous studies, the proposed approach outperformed all previous studies when dataset 2 and part of dataset 1 were used. For the complete dataset 1, the proposed model performed very well. | http://dx.doi.org/10.12785/ijcds/150119 | Intelligent Tutoring System; Convolutional Neural Network; Machine Learning | Fatema Abdullatif Alshaikh and Nabil Hewahi (UOB, Bahrain) |
| 20 | 1570892866 | Internet of Things Based Low-Cost Health Screening and Mask Recognition System | Face mask can reduce the spread of several virus as well as corona virus and help people to save many lives. Many people are not serious about wearing face mask. This work results a system that can detect face mask of a person with CNN. The system can identify anyone without face mask. An IoT based module is integrated with the system to monitor people's temperature, blood pulse and oxygen level. Which can collect all the sensor data and send the data to a cloud database. From the cloud server the mask condition and health condition can be monitored. The collected data can also be used for future analysis. The proposed low-cost system worked properly with a good success rate. | http://dx.doi.org/10.12785/ijcds/150120 | Face-mask detection; Internet of Things; E-health; Health Screening; Machine Learning; CNN | Tajim Md. Niamat Ullah Akhund (Daffodil International University, Bangladesh & Saga University, Japan); Nishat Tasnim Newaz and M. Mesbahuddin Sarker (Jahangirnagar University, Bangladesh) |
| 21 | 1570893242 | Cyclic Olefin Copolymer Based Photonic Crystal Fiber as Single Mode THz Waveguide | For global broadband networks, fiber optic systems are vital components of the telecommunications infrastructure. Signal transmission with a broad bandwidth and minimal delay is a fundamental need for modern applications. Owing to its low transmission losses, the adoption of optical fiber is growing exponentially. Researchers are exerting significant effort to improve the transmission efficiency of fibers with extremely low loss. A suspended porous core Cyclic Olefin Copolymer (COC) based fiber model is proposed in this work. It is exceptionally transparent and crystalline plastic. This model is challenging to build a waveguide with a low loss for THz wave propagation since most polymers are particularly absorbent. The proposed fiber model also has an optimized core diameter. In this analysis, a fiber with a low EML of 0.039456 cm-1, a marginal due to confinement loss is 2.264 × 10-5 dBcm-1, low dispersion of 0.8594 psTHcm-1 and V-parameter is 1.142 for 300 core diameter and at 1 THz working frequency. Additional important propagation properties, such as the power fraction of the proposed fiber have also been thoroughly investigated. | http://dx.doi.org/10.12785/ijcds/150121 | Birefringence; Effective material loss; Photonic crystal fiber; THz Guidance; Cyclic olefin copolymer | Anika Tun Naziba (American International University-Bangladesh, Bangladesh); Manika Tun Nafisa (Southern Polytechnic College of Engineering and Engineering Technology, USA); Mohammad Nasir Uddin (American International University-Bangladesh & IEEE Executive Committee Bangladesh Section (EAC), Bangladesh) |
| 22 | 1570914887 | Enhanced approach to generate One Time Password (OTP) using Quantum True Random Number Generator (QTRNG) | In secure IT systems, a One Time Password (OTP) is used. They can only be used once for a transaction or a login session. It is usually used for multi-factor authentication and is one of the security services based on human responses. In the field of cryptography, the creation of pseudo-random bitstreams is similar to how one-time passwords (OTPs) work. In modern computers, the security of cryptographic procedures is greatly enhanced by the use of random numbers. The use of a Quantum True Random Number Generator (QTRNG) makes it possible to replace periodic sequences that appear random with real random data. A QTRNG is used in this research on a real Quantum Computational Device (QCD) in a local environment. The IBM Quantum Experience Qiskit platform is used to build a random number generator. Qiskit is an open source platform that provides users to work with actual quantum devices for simulating quantum algorithms, different gates and circuits on IBM Quantum Experience or on simulators. The random number is investigated with different qubits (quantum bits) to generate random bits that can be used for One Time Password Generation. | http://dx.doi.org/10.12785/ijcds/150122 | Hadamard Gate; NIST Statistical Test; OTP; Qiskit; QTRNG | Riddhi B. Prajapati (Graduate School of Engineering and Technology, Gujarat Technological University, India); Shailesh Panchal (Gujarat Technological University, India) |
| 23 | 1570897196 | An Efficient Tamper Detection and Recovery Scheme for Attacked Speech Signal | Data security is an essential communication issue, where the resistivity to data tampering, corruption and attacks forms a critical problem. This paper proposes a robust approach for tamper detection in a speech signal and recovering the destroyed parts of the attacked signal from the embedded spare parts using a specific strategy. A watermarking method is suggested to embed authentication codes, synchronization identifiers, and spare part data. The authentication codes are generated using the Singular Value Decomposition (SVD) method, while G.723.1 speech CODEC generates the spare part data. The proposed strategy considers attacks such as muting and replacement. According to the experiments, the quality of the watermarked and retrieved signals has improved. The average signal-to-noise ratio is above 70 dB for the watermarked signal, and the normalized correlation for the retrieved signal is close to attacked signals. Additionally, this approach achieves less distortion in the watermarked signal quality measured by log spectrum distortion, which is about 0.0072. The proposed approach is compared to the related methods, where it achieves the best results, including high robustness against different types of attacks. | http://dx.doi.org/10.12785/ijcds/150123 | Tamper Detection,; Signal Recovery; Speech Watermarking; Singular Value Decomposition; Authentication Code; Speech Signal Attack | Younis Mohammed Jalil (University of Kerbala & College of Engineering, Iraq); Haider Shahadi and Hameed Rasool Farhan (University of Kerbala, Iraq) |
| 24 | 1570897366 | A comprehensive literature review on air-written online handwritten recognition | The COVID-19 pandemic has shifted ordinary life into a digital platform. Individuals rapidly use digital or online media to avoid touch-based platforms' impact. In this case, touchless technology is primarily used in operation theatres, online education systems, ticket counters, etc. However, the online numerals or characters written in the air are challenging tasks to recognize efficiently due to the writing styles of every individual. Therefore, the present research focuses on recent work on online air-written handwritten recognition. Almost a seventy plus standard articles from various journal/conference databases on handwriting recognition are considered for the present review. In addition, the generalized architecture of handwriting recognition is also provided for new research. It is observed that some online handwritten databases for some languages are freely available for research purposes. However, the online real-time air-written numeral dataset is private for Devanagari and English. Therefore, designing and developing a standard dataset for character or numeral recognition written in the air is suggested. The earlier studies achieved satisfactory results on air-writing multilingual numerals collected via or without sensors using advanced machine learning and deep learning methods. The convolutional neural network (CNN) has provided the best accuracy for several languages excluding Devanagari. | http://dx.doi.org/10.12785/ijcds/150124 | Air-writing data; online handwriting recognition; air written recognition; convolutional neural network; Devanagari/English numeral dataset | Meenal Jabde (Vishwanath Karad MIT World Peace University, Pune, India); Chandrashekhar Himmatrao Patil (Vishwanath Karad MIT World Peace University, India); Amol D. Vibhute (Symbiosis International Deemed University Pune, India); Shankar Mali (Vishwanath Karad MIT World Peace University, India) |
| 25 | 1570900563 | Utility and acceptability of AI-enabled chatbots on the online customer journey | Retailers use artificial intelligence (AI) to serve customers better. This study examines the customer's awareness level of chatbots are used in the e-retailing sector to serve as the interaction mode among the customers and brands along the customer journey. Chatbots act as automated conversational agents capable of understanding customers, providing them product knowledge, and promoting behavioral change. Chatbots provide first-level support atomized and cost-effective because a human employee can not answer all the customer questions round the clock. The authors in the study have surveyed to analyze chatbot's perceived utility and acceptability based on the customer's awareness level in retailing using the SPSS software. The online questionnaire was circulated via Google forms (N-204) promoted on social media. This study's novelty is to determine how the online customer journey is enhanced due to the application of AI-enabled chatbots in e-retailing. The study focuses on determining the customer's perspective on AI-enabled chatbot's utility and acceptance in the online customer journey. The quantitative investigation revealed that a significant majority of respondents were apprehensive about employing AI and online-retail (e-retail) chatbots, primarily due to concerns about their accuracy and security, but would like to seek the advice of chatbots for informative purposes. | http://dx.doi.org/10.12785/ijcds/150125 | Artificial intelligence (AI); chatbot; e-retail; perceived utility; acceptability; customer journey | Jyoti Rana (School of Management & Commerce, K. R. Mangalam University, Gurugram, India); Ruchi Jain (Amity University Uttar Pradesh India, India); Vibha Nehra (Amity University, Uttar Pradesh, India) |
| 26 | 1570900846 | Design and Implementation of an Automated Indoor Hydroponic Farming System based on the Internet of Things | Urban farming has been growing in popularity to help secure food needs in urban areas. Hydroponics is one of the methods to grow crops without soil media being an option for urban farming. However, hydroponic farming has its challenges as farmers must carefully monitor the environmental conditions of the plants regularly and adjust the nutrient solution and water circulation based on the environmental conditions. This paper describes designing and implementing an IoT-enabled Hydroponic farming system to monitor the environmental condition of plants and control the nutrient supply plants. Our system is built from sensors, actuators, a micro-controller unit (Arduino), and a single-board computer (Raspberry pi) attached to the hydroponic system. The system can monitor the environmental condition of hydroponics through sensors such as temperature, humidity, pH sensors, and Total Dissolve Solid (TDS) sensors and control water pumps to circulate the nutrients to the plant. The Raspberry pi acts as a Message Queuing Telemetry Transport (MQTT) broker for distributing the data from sensors to subscribers. It also controls the nutrient pump to adjust nutrient solution. We use Node-RED installed in the Raspberry pi to build and connect the system to hardware devices. Users can monitor the environmental condition of plants in hydroponics through a web browser on their smartphone and laptop. This IoT-based indoor hydroponic system can automate the delivery of nutrients and water to plants, ensuring they receive the optimal amount at the right time | http://dx.doi.org/10.12785/ijcds/150126 | Internet of Things; Hydroponic; Sensors; MQTT; Node-Red | Muhammad Niswar (Universitas Hasanuddin, Indonesia) |
| 27 | 1570909860 | Use Word Cloud Image Of Web Page Text Content On Convolutional Neural Network (CNN) For Classification of Web Pages | In today's environment, people can easily use the internet to find information by visiting web pages. Most people like to visit web pages that offer games and videos to watch online. People who spend a lot of time on web pages like these can become addicted to the internet and it can have a bad effect on them. Access to web pages that offer games and streaming videos needs to be limited to stop people from being addicted to the internet. It needs a tool that can classify web pages category based on its content. Due to lack of matrix representation that unable to handle long web page text content, this study uses a technique which is word cloud image to visualize the words that has been extracted from the text content web page after performing data pre-processing. The most popular words from the text content web page are displayed in big size and appear in center of the word cloud image. The most popular words are the words that frequently appear in the text content web page, and it related to describe what the web page content is about. The Convolutional Neural Network (CNN) identifies the pattern of words displayed in the central areas of the word cloud image to classify the category that the web page belongs to. The proposed model for classifying web pages has an accuracy of 0.86. The proposed model can be used, for example, by the institution to set rules and limit the usage of the internet for the users to surf the web pages that offer games and streaming videos. It will be one of the ways to prevent users from getting internet addiction. | http://dx.doi.org/10.12785/ijcds/150127 | Web page classification; document representation; word cloud image; deep learning; Convolutional Neural Network | Siti Hawa Apandi, Jamaludin Sallim and Rozlina Mohamed (Universiti Malaysia Pahang, Malaysia) |
| 28 | 1570910216 | Verse-Based Emotion Analysis of Bengali Music from Lyrics Using Machine Learning and Neural Network Classifiers | Throughout the decades, music has been involved with humans' emotions and inner imaginations. Along with the overgrowing demands for the overhead reduction of human-aided processes, the urge for automation in the linguistic domain has become a preponderance. Research related to Natural Language Processing (NLP), have been flourished well enough in the case of English and other contemporary modern languages. Although Bengali is an enriched multipurpose language with so wide varieties and precious dialects, very few researches have been conducted on emotion recognition in Bengali music. Besides, the Bengali music archive is blessed with pioneering works of some world-class writers, and in addition to that, modern music enthusiasts' versatile music predilection has made the task of automation in sentiment analysis quite interesting and challenging. In this study, we introduce a verse-based emotion analysis system for Bengali songs that is able to identify certain emotions from textual data. To appraise the relevant results of our resorted approach, we used several Machine Learning (ML) and Neural Network (NN) models. Eventually, we were able to achieve the best accuracy of 80% and 65% using Bidirectional Encoder Representations from Transformers (BERT) model for both three and two emotion classes. | http://dx.doi.org/10.12785/ijcds/150128 | Natural Language Processing; Music Information Retrieval (MIR); Sentiment Analysis; Multi-Class Classification; Machine Learning; Neural Network | Maraz Mia (Tennessee Technological University, USA); Pulock Das (Shahjalal University of Science and Technology & TallyKhata, Bangladesh); Ahsan Habib (Shahjalal University of Science and Technology, Bangladesh) |
| 29 | 1570920677 | A Survey on Leveraging Blockchain for IoT Security | IoT is a very fast-growing technology. IoT can be defined as interconnected small smart devices linked over the Internet to communicate with each other to perform meaningful action. There are major concerns regarding the security of data being produced from millions of devices in the IoT system. Different security concerns in various IoT system tiers have been covered in this study. IoT security concerns can be reduced by using Blockchain technology, which is a decentralized distributed ledger with several Blockchain potentials, including Persistence, transparency, verifiability, Encryption, and operationally strong The paper reviews whether they make a good fit along with certain challenges of Blockchain that should be examined while integrating it with IoT for resolving various security issues. | http://dx.doi.org/10.12785/ijcds/150129 | Internet of Things; Blockchain; Security; Ethereum; Integration | Nishant Suresh Sanghani (Gujarat Technological University, India); Bhavesh Borisaniya (Shantilal Shah Engineering College, Bhavnagar, India) |
| 30 | 1570911792 | Overcoming the Challenge of Cyberbullying Detection in Images: A Deep Learning Approach with Image Captioning and OCR Integration | Cyberbullying is a serious concern in today's digital age. The rapid increase in the use of social media platforms has made cyberbullying even more prevalent. The form of cyberbullying has also evolved with time. In the era of Web 1.0, cyberbullying was limited to text-based data, but with the advent of Web 2.0 and 3.0, it has expanded to images and multi-modal data. Detecting cyberbullying in text-based data is relatively easy as various natural language processing techniques (NLP) can be used to identify offensive language and sentiment. However, detecting cyberbullying in image-based data is a major challenge as images do not have a clear textual representation. Hence, bullies often try to bypass existing cyberbullying detection techniques by using images and multi-modal data. We proposed a deep learning technique named as CNBD-Combinational Network for Bullying Detection (CNBD), which is a combination of two networks: Binary Encoder Image Transformer (BEiT) and Multi-Layer Perceptron (MLP) network. To improve the performance of the CNBD, we supplied two additional input factors to the CNBD using concepts called Image Captioning(IC) and OCR (Optical Character Recognition) to extract text overlayed on the images. The experimental results proved the two additional factors gave an advantage to the CNBD technique in terms of accuracy, precision, and recall. | http://dx.doi.org/10.12785/ijcds/150130 | Cyberbullying; Social media; Multi-Layer Perceptron; Deep Learning; Image Captioning | Subbaraju Pericherla and Ilavarasan E (Pondicherry Technological University, India) |
| 31 | 1570891225 | HyARBAC: a new Hybrid Access Control Model for Cloud Computing | One of the most important digital processes and storage technologies in recent years is cloud computing. Nevertheless, the cloud is faced to data security issue due to its vulnerability to weak access control. In fact, access control guarantees accessing cloud data and resources only to authorized users, then the unauthorized users are detected and prevented from retrieving these resources. There are many models of access control that have been proposed and used in the cloud computing domain, such as Attribute Based Access Control (ABAC) and Role Based Access Control (RBAC), however, these schemes suffered from many limitations, such as their difficulty using contextual information, such as time, location of user, type of device or the use of attributes residing in multiple and disparate locations, leading the performance of such approaches to be extremely low. To deal with these problems, we present in this article a novel model of access control applied to Cloud computing, called Hybrid attribute and role-based access control for cloud computing (HyARBAC). This proposal combines ABAC and RBAC provide a flexible and accurate access control, which takes into account environmental information when controlling access and decreases the administrator intervention to manage and control this task. In addition, this model is reinforced by the integration of a moving target defense mechanism (MTD) considered as an extra layer of security for deterring current and upcoming threats. Trying to compromise the original attributes and their corresponding mechanisms of our authorization policies. After an experimental study applied on a healthcare database and comparison against ABAC and RBAC according to several performance features, the effectiveness of HyARBAC was proved to enhance access control. | http://dx.doi.org/10.12785/ijcds/150131 | RBAC; ABAC; Cloud computing; Moved Target Defense; Access control | Okba Houhou (University of Biskra, Algeria); Salim Bitam (University of Biskra & LESIA Laboratory, Algeria); Hamida Ammar (University of Biskra, Algeria) |
| 32 | 1570892077 | A framework for Multimedia Data Mining using Transformer based intelligent DNN model architecture | Multimedia data mining plays a crucial role in various fields, such as image and video analysis, natural language processing, and recommendation systems. Multimedia data refers to any form of data that involves multiple modes of communication, such as text, images, audio, and video. To effectively mine valuable insights from multimedia data, a new framework is proposed in this paper that employs a transformerbased intelligent deep neural network (DNN) model architecture. The framework includes an extensive data preprocessing step that involves obtaining multimedia data from internet searches and removing duplicates to ensure that each image is unique. The proposed transformerbased intelligent DNN model architecture processes the multimedia data in a hierarchical manner and utilizes shifted windows to achieve high accuracy in image classification task. The exploited dataset details are provided in the experimental evaluation section. Experimental results show that the proposed framework outperforms existing multimedia data mining methods in terms of accuracy and efficiency. This framework provides valuable insights that can be used in various applications, including content based image retrieval, sentiment analysis, and automated captioning. | http://dx.doi.org/10.12785/ijcds/150132 | Multimedia data mining; Images classification; Transformers; Deep neural nets; Computational intelligence; Statistical performance metrics | Ravi M (Jawaharlal Nehru Technological University, Hyderabad, India); Ekambaram Naidu M (SRK Institute of Technology Vijayawada, India); Narsimha Gugulotu (India) |
| 33 | 1570894267 | Physical Internet Based Ontology for Supporting Traceability in Logistic IoT | Current supply chain management (SCM) requires the control of physical and information flows in order to satisfy the customer, i.e. deliver the right product to the customer at the right place, at the right time, at the right price and at the lowest cost. SCM is inseparable from traceability which makes reliable the said flows, accelerates the transmission of information on these flows, allows to access a detailed knowledge of the movements, and makes the flows visible. In order to streamline and monitor, if possible in real time and permanently these logistical processes, we propose the design and implementation of an Ontology-based traceability system based on an architectural model for the physical Internet using computing resources such as Cloud computing, Fog computing and Internet of Things (IoT) to achieve efficiency and sustainability goals. To evaluate our system, we were able to carry out all the queries that the user can express whether he is a customer, a supplier or a manager. | http://dx.doi.org/10.12785/ijcds/150133 | Supply Chain Management; Logistics; Traceability Systems; Physical Internet; Internet of Things; Cloud and Fog Computing | Mohammed Frendi (Oran 1 University, Algeria); Bakhta Nachet and Abdelkader Adla (Oran1 University, Algeria) |
| 34 | 1570897616 | A Future Trading System using Ensemble Deep Learning | Stock price prediction is a challenging research problem due to various dependencies like news headlines, Twitter, microblogs, and price history, stock price prediction. The news headlines and the price history are the most significant features for predicting the stock price, which helps investors make sell or buy decisions to maximize the profit. The manual analysis of news and price history information is impossible; hence automation is required to summarise the available information for wise trading decisions. We propose an ingenious Future Trading system with Ensemble Deep Learning (FTSEDL) to improve the correctness of the prediction.The objective is to develop FTSEDL that improves the Deep learning predictive model with a new context-based clustering, CbCA, and information fusion for an accurate stock value forecast. The present research experimented with six stocks with news headlines and price history collected from 8th August 2016 to 5th August 2021, i.e., five years. The proposed model achieved an accuracy of 91.54%, MAE of 3.09, RMSE of 13.14, and MAPE of 0.02, which outperformed the models in the literature. The FTSEDL has many advantages when compared with recent literature. In addition, there is vast scope to apply this concept in real-time. | http://dx.doi.org/10.12785/ijcds/150134 | Artificial Intelligence; Feature Extraction; Stock; Fusion; Clustering | Ssssn Usha Devi N (National Institute of Technology Tiruchirappalli, India & University College of Engineering Kakinada, India); Mohan R (National Institute of Technology Tiruchirappalli, India) |
| 35 | 1570898252 | Region Detection Technique Using Image Subtraction and Pixel Expansion Cue For Obstacle Detection System On Small - Sized UAV | This research paper is about method of detection of free region and obstacle region by combining image segmentation and frame subtraction method. The application will be further study to be used by Unmanned Aerial Vehicles (UAVs). This method intends to minimize the weight of UAV by avoiding heavy sensors. Objective of this research is to utilize the pixel expansion of object to find free region. K-means segmentation will be used to separate the interest area from the background. Then, segmented image frame will be subtracted and then divided into several grids and the amount of subtracted pixel that has been set into yellow and black color of each grid will be calculated with respect of distance given. The expansion of pixel will be detected as, the distance between image frame coming closer, number of obstacle pixel will be higher. The application of simple LIDAR that emits single ray to frontal obstacle will initiate the camera to capture image frame to be further analyze. Experiment was carried out in close environment with different cases and total of 100 images has been captured that consist of texture obstacle, texture-less obstacle and multiple obstacles. The findings showed bearable result as the free region detection is 88.0% for texture obstacle and up to 84.0% of free region were successfully detected for texture-less obstacle. Due to lack of cues and texture in texture-less object, the algorithm had difficulties to detect the center of object that expand in static form. | http://dx.doi.org/10.12785/ijcds/150135 | Image segmentation; Obstacle detection; Monocular camera; Pixel expansion; Free region detection; Small -Sized UAV | Muhd Wafi Abdul Aziz (Universiti Tun Hussein Onn Malaysia (UTHM) & Research Centre for Unmanned Vehicle (ReCUV), Malaysia); Muhammad Faiz Bin Ramli (Universiti Tun Hussein Onn Malaysia, Malaysia) |
| 36 | 1570901820 | Optimizing The Maximum Power of Photovoltaic System Using Modified Incremental Conductance Algorithm Operating Under Varying Dynamic Climatic Conditions | In this study of research a hybrid MPPT consisting of a well-known mechanism defined as incremental conductance algorithm (INC) is being adapted to operate along with fuzzy logic controller (FLC). By adapting and incorporating these two techniques, the maximum power point (MPP) can be found and tracked at different levels of solar irradiances. Additionally, it will work on generating a modulated signal to allow the adjustment of the duty ratio cycle of the DC-DC converter. This is a requirement in this research to speed up the MPPT tracking. Simulation analysis is performed to verify the accuracy of tracking levels at different irradiance levels of our proposed approach. The validated results revealed that our hybrid system was able to track the maximum power (Pmax) and locate the MPP under different solar irradiances and decrease the variation at the MPP range. Furthermore, the main scheme proposes an MPPT algorithm relies on utilizing an adaptive variable step size INC based fuzzy logic controller. This system and at each iteration it changes the step size according to the location of the operating point with respect to the maximum power point by reducing the above trade-offs through increasing the convergence speed without affecting the compromising of accuracy. | http://dx.doi.org/10.12785/ijcds/150136 | Fuzzy logic; adaptive step size; control system; photovoltaic | I M Elzein (University of Doha for Science and Technology, Qatar); Moustafa Mohamad Kurdi (City University & Engineering and Information Technology, Lebanon); Yasen Harrye (University of Doha for Science and Technology, Qatar) |
| 37 | 1570902581 | Forecasting the Consumer Price Index: A Comparative Study of Machine Learning Methods | The Consumer Price Index is an indicator of inflation and is tracked by many government and economic agencies to make decisions of major importance. In this paper, we forecast the Saudi Arabian Consumer Price Index with machine learning (ML) methods. We compare the performances of Decision Tree, kNN, Linear Regression, Neural Networks, Random Forest, and SVM, all applied to the 2013-2020 Saudi CPI dataset. Experiments were conducted to vary the training and testing sets, optimize the machine learning parameters, and improve the MSE and R2 metrics. The CPI predictions of these methods were also compared to the 2021-2024 International Monetary Fund forecast and the actual 2021-2024 CPIs. The results indicate that the multilayer perceptron neural network outperforms the other methods, is nearest to the actual CPI, and may be used to forecast the CPI for up to 3 years from the latest CPI data in the training dataset. | http://dx.doi.org/10.12785/ijcds/150137 | Consumer Price Index,; Forecasting; Machine learning; Data Science | Fadi Sibai (Gulf University for Science and Technology, Kuwait); Ali A. El-Moursy (University of Sharjah, United Arab Emirates); Ahmad Sibai (University of South Florida, USA) |
| 38 | 1570905786 | Improving the Prediction Accuracy of MRI Brain Tumor Detection and Segmentation | Brain tumors were the most common kind of tumor in humans. Brain tumor scan be detected from various imaging technologies. The proposed research work strives to improve the prediction accuracy of brain tumor detection and segmentation from MRI of human head scans by using a novel activation function E-Tanh. The role of activation functions is to perform computations and make decisions in artificial neural networks (ANN). We developed three ANN models for brain tumor detection by modifying the hidden layers. We have trained these ANN models using the E-Tanh activation function and evaluated their performance. This novel activation function achieved 98% prediction accuracy for the MRI brain tumor image detection neural network model, which was higher than the existing activation functions. We also have segmented brain tumors from the BraTS2020 dataset by using this activation function in U-Net-based architecture. We attained dice scores of 83%, 95%, and 85% for the whole, core, and enhancing tumors, which are significantly higher than the ReLU activation function. | http://dx.doi.org/10.12785/ijcds/150138 | Activation Functions,; Artificial Neural Networks,; MRI, | Padmapriya Thiyagarajan and Chandrakumar T (Thiagarajar College of Engineering, India); Kalaiselvi T (Assistant Professor, India) |
| 39 | 1570906756 | Exploring Factors Affecting Citizens' Acceptance to Use E-Participation in Malaysian Local Governments Through an Extended UTAUT Model | E-participation platforms in local government are now recognized as a digitalization strategy for enriching the democratization culture in citizen-government relations, especially at the grassroots level. This article explores the current understanding of the relationship between Malaysian citizens' acceptance and use of e-participation platforms provided by their respective local governments, which also will oversee the moderating effect analysis of demographics. This study framework integrates with slight extensions to the original Unified Theory of Acceptance and Use of Technology (UTAUT) model. The study uses a quantitative methodology involving a survey questionnaire with convenience sampling applied to respondents comprised of 484 local Malaysian citizens via offline and online methods. IBM SPSS Statistics Version 27 software will be used for data analysis, comprising Pearson Correlation for correlation analysis and structural equation modeling analysis via SmartPLS 3.0 software for moderating effect analysis. The findings show that all citizens' acceptance factors correlate with using e-participation. In the meantime, all 20 demographics' moderating effects were tested, but no significant relationship was found. This study is significant in that it revisits several relevant literature reviews to enrich the development of the research framework and provide a necessary foundation for understanding the use of e-participation within Malaysian local governments. | http://dx.doi.org/10.12785/ijcds/150139 | Citizens' acceptance; Public participation; E-participation; Malaysian local governments; UTAUT | Mohamad Nur Salam Man and Halimah Abdul Manaf (Universiti Utara Malaysia, Malaysia) |
| 40 | 1570907040 | Lightweight Neural Network Design for Real-time Human Tracking with Visual Modality | We proposed a lightweight neural network architecture model that focuses on efficient computation when doing face detection in real-time with limited system resources, in image processing time to be exact. The concept is to implement layers and neurons as minimum as possible to reduce processing time and computing resource used in either convolutional layer or fully connected layer. This research is conducted because current neural network technology does not consider real-time detection scenario needs, such as tracking an object using camera. The result of proposed neural network implementation is an application that captures video from camera and generates boundary box that contains human face, with 0.117 seconds processing time each data, 96.735% accuracy, and 0.1219 error rate. The proposed model used 169.3 MB RAM and taking up to 1.186% CPU processing. Although does not have best accuracy and error rate, the proposed method does have faster processing time and lower usage of system resources in human face detection. This will allow computers with lower specification to use this model for face detection, especially in human tracking. This research also provides the concept of convolutional neural network, object tracking, and face detection that would be the base of create the proposed lightweight neural network. | https://dx.doi.org/10.12785/ijcds/150140 | neural network; face detection; object tracking; human tracking | Christopher Limawan and Benfano Soewito (Bina Nusantara University, Indonesia) |
| 41 | 1570907182 | Emotion Detection using Stack Auto Encoder, Deep Learning and LSTM Model | There is an increasing need for distress emotion acknowledgement and channelling the same is very important. There is an increasing need for machines to understand human and their complex emotions deeply. This research describes a unique framework for emotion detection that helps brain-computer interface (BCI) to understand human emotions and brain complexity and working using multi-channel electroencephalograms (EEG). We have employed two datasets in our work, Database for Emotion Recognition Through EEG and ECG Signals from Wireless Low-Cost Off-the-Shelf Devices (DREAMER) and dataset for emotion analysis using EEG, physiological and video signals (DEAP) for LSTM model implementation and validation respectively. A linear model is considered for EEG signal mixing and an emotion timing model comprise the framework. By harnessing the context correlations of the EEG feature sequences, our proposed methodology efficiently separates the EEG source signals from the retrieved EEG signals and enhances classification accuracy rate. Also, stress-bins were created for each user to detect his level of stress or calm after being exposed to external stimuli. DEAP dataset using LSTM framework was enforced for emotion recognition, and mean recognition accuracy using area under curve as evaluation matrix for valance and arousal was 82.02% and 76.52% respectively, validating competence of framework. Novelty of our work is it improved competency in feature extraction, use of context correlations increasing accuracy and use of spatio-temporal features in the proposed model framework. | https://dx.doi.org/10.12785/ijcds/150141 | Sparse Auto Encoder; electroencephalogram (EEG); OpenBCI; LSTM Model; Emotion Detection | Disha Shah and Rashmi Rane (Dr Vishwanath Karad MIT World Peace University Pune, India); Shakti Kinger (Vishwanath Karad MIT World Peace University, Pune, India) |
| 42 | 1570908116 | Multi-Faceted Job Scheduling Optimization using Q-learning with ABC in Cloud Environment | Resource allocation is the most challenging and common problem, particularly in the cloud service model in Infrastructure as a Service. The issue of load balancing is so harmful that irregular load balancing may result in a structure smash. Adopting a suitable access plan and allowing the system to spread work among all existing resources leads to utilizing Virtual Machines (VMs) appropriately. To get enhanced results from the Artificial Bee Colony Algorithm (ABC), a reinforcement learning technique Q-learning is combined using multifaceted job scheduling optimization based on ABC(QMFOABC) has been proposed. The proposed approach improves resource utilization and scheduling created on resource, cost, and makespan. The efficiency of the suggested strategy was evaluated in datasets synthetic workload, google cloud jobs (GoCJ), and Random by using CloudSim to the remaining scheduling strategies for load matching like Max-min, Multifaceted Cuckoo Search (MFCS), Multifaceted Particle Swarm Optimization (MFPSO), Q-learning, Heuristic job scheduling with Artificial Bee Colony approach with Largest Job First algorithm (HABC_LJF), First come first serve (FCFS). According to the findings of the experiments, the algorithm that employed the QMFOABC method has better results in resource utilization, throughput, cost, and makespan. Compared to Max-Min (82.31%), MOPSO (35.62%), HABC_LJF (21.65%), Q-Learning (11.72%), VTO-QABC_FCFS(5.87%), and VTO-ABC_LJF (5.86%) shorter time than MOCS, a considerable improvement is found. | http://dx.doi.org/10.12785/ijcds/150142 | Cloud computing; multifaceted job; scheduling; Q-learning; virtual machines; ABC | Neeraj Kumar Pandey (Graphic Era Deemed to be University, India); Sanjeev Sharma (School of Computing DIT University Dehradun, India) |
| 43 | 1570911627 | Solve Cocktail Party Problem Based on Hybrid Method | One of the most intractable issues in contemporary Digital Signal Processing (DSP), particularly with regards to Blind Source Separation (BSS) Methods, is known as the Cocktail Party dilemma. This problem suppose there are many sensors record many signals at same time to produce many mixed signals. To solve this problem, one of an important methods used for this purpose is an Independent Component Analysis (ICA) Method. This method abbreviates in How separate mixed signals without any pre-knowledge about the mixing signals?. It treats on the statistical features of a mixing signals. This work introduces new novel method to solve the cocktail party problem, by using hybrid method from the Quantum Particle Swarm Optimization (QPSO) method and the Bell- Sejnowski neural method to enhance the performance of the ICA. In addition, the proposed method uses the Negentropy function to be the objective function of the optimization process. The proposed algorithm has been implemented on two cases of three really signals, with 8-KHz frequencies. The results of the separating process measured in two directions: Firstly by comparing the results with other methods as Particle Swarm Optimization and the Quantum Particle Swarm Optimization, where the results appear that the proposed method appears very high results than other methods. Secondly, by using standard metrics as SNR, SDR, and AVCC. | http://dx.doi.org/10.12785/ijcds/150143 | Cocktail-Party Problem; Blind Source Separation (BSS); Independent Component Analysis (ICA); Quantum Particle Swarm Optimization (QPSO); Bell-Sejnowski (InfoMax) method | Hussein Mohammed Salman (University of Babylon, Iraq); Ali Kadhum M. Al-Qurabat (University of Babylon & College of Science for Women, Iraq); Abd alnasir Riyadh Finjan (Supreme Commission for Hajj and Umrah, Iraq) |
| 44 | 1570911928 | An Efficient Spam and Phishing Email Filtering Approach using Deep Learning and Bio-inspired Particle Swarm Optimization | The exponential growth of spam and phishing emails significantly threatens users' privacy, security, and productivity. This research shows a deep learning model that demonstrated improved performance compared to similar state-of-the-art studies. Deep learning neural networks, with properly fine-tuned settings, can accurately sort emails into three categories: normal (ham), unwanted (spam), and deceptive (phishing). We use two popular techniques, Principal Component Analysis (PCA) and Particle Swarm Optimization (PSO), to select the essential features for email classification. Test results show that the PSO method outperforms the others, achieving an impressive 99.35% accuracy rate. The findings highlight the potential of deep learning algorithms for effective email filtering, helping to detect and reduce spam and phishing. Additionally, the study emphasizes the importance of feature selection in enhancing the performance of deep learning models for sorting emails as normal (ham), spam, or phishing. | http://dx.doi.org/10.12785/ijcds/150144 | Spam email; Phishing email; Artificial neural network; Particle swarm Optimization | Santosh Kumar Birthriya (National Institute of Technology Kurukshetra India, India); Priyanka Ahlawat (National Institute of Technology ( Institute of National Importance), India); Ankit Kumar Jain (National Institute of Technology, Kurukshetra, India) |
| 45 | 1570927848 | Optimized Deep Neural Networks Using Sparrow Search Algorithms for Hate Speech Detection | Deep learning has widespread use in various domains, including computer vision, audio processing, and natural language processing. The hyperparameters of deep learning algorithms have a significant impact on the performance of these algorithms. However, it can be challenging to calculate the hyperparameters of complicated machine learning models like deep neural networks due to the nature of the models. This research suggested a strategy for hyperparameter optimization utilizing the Long Short-Term Memory with Sparrow Search Algorithm (LSTM-SSA) model. The model that has been presented uses a deep neural network, which can recognize and classify instances of hate speech as either hate speech or neither. Experiments are conducted to validate the suggested technique in both straightforward and intricate network environments. The LSTM-SSA model is validated using a dataset consisting of hate speech, and an experimental investigation into the model's sensitivity, accuracy, and specificity is carried out. The outcomes of the experiments demonstrated that the suggested model might be improved upon, as it had an accuracy of 0.936. | http://dx.doi.org/10.12785/ijcds/150145 | Hate Speech; LSTM; NLP; social media; sparrow search algorithm | Ashwini Kumar (Graphic Era Deemed to Be University, Dehradun, India); Santosh Kumar (Graphic Era University, Dehradun, India) |
| 46 | 1570923393 | Deep Neural Network-based Current and Voltage Prediction Models for Digital Measuring Unit of Capacitive Resistivity Underground Imaging Transmitter Subsystem | Real-time monitoring of output electrical parameters of the transmitted signals in a capacitive resistivity underground imaging system is necessary because these are significant in the calculation of underground resistivity, however, machine learning has not yet been applied in this application to improve the accuracy of measurement. This study aims to develop and select the best prediction models that can be implemented for a digital measuring unit suitable for capacitive resistivity underground imaging. Three deep neural network models namely Elman recurrent neural network (ERNN), long short-term memory (LSTM), and gated recurrent unit (GRU) were explored to build prediction models for the current and voltage of the transmitter circuit. The prediction models' performance was assessed using mean squared error (MSE), which is reduced to its absolute lowest value. The result shows that the best-trained models for current and voltage prediction are the ERNN models with configurations of 900-600-500 hidden neurons network with training MSE of 9.82 × 10^-9 and the configured 1300-1000-900 hidden neurons with training MSE of 0.465, respectively. With the help of the prediction models, it would be possible to measure current and voltage output accurately, allowing simultaneous data acquisition while avoiding the need for a separate measuring device. | http://dx.doi.org/10.12785/ijcds/150146 | transmitter antenna; digital measuring circuit; recurrent neural network; long short-term memory; gated recurrent unit; underground imaging | Jonah Jahara G Baun, Adrian Genevie G. Janairo, Ronnie Concepcion II, Kate G Francisco, Mike Louie Cruz Enriquez, R-jay S Relano, Edwin Sybingco, Argel Bandala and Ryan Rhay P. Vicerra (De La Salle University, Philippines) |
| 47 | 1570912771 | A Hybrid Approach Based on Boosting Algorithm for Effective Android Malware Detection | Mobile phones are being used much more often and play a crucial role in our everyday lives. These gadgets hold a tonne of personal information and provide a variety of functions and services. Mobile devices have become indispensable for those who utilise technology and communication since they are practical and effective. The complexity and sophistication of mobile apps is rising as a result of the diverse demands of consumers, and both hardware and software developers are working to increase battery life and performance of mobile devices. The market offers a variety of mobile operating systems, each with a unique platform and market share. However, mobile systems are vulnerable to virus assaults just like any other type of information system.The development of hardware technologies has increased the complexity and performance of mobile apps.Mobile devices come with a lot of advantages, but they also have a lot of drawbacks. The need to strike a balance between performance and battery life is one such difficulty. The complexity and power requirements of mobile apps are increasing as they grow more demanding, which reduces battery life. Additionally, the danger of security breaches and data theft rises with continued usage of mobile devices. Malicious actors may use mobile system flaws to obtain sensitive data, such as login passwords, financial information, and personal information.Mobile device makers and app developers are constantly changing their software to enhance security and performance in order to solve these issues. For instance, to safeguard user data, many mobile operating systems now have built-in security measures like firewalls, encryption, and two-factor authentication. To maximise battery life, designers are also creating mobile apps that utilise resources wisely and require less power.In conclusion, mobile devices are becoming an essential component of our life, providing efficiency and ease in technology and communication. As with any technology, they come with a number of drawbacks, such as security issues and battery life. Manufacturers and designers must prioritise security and performance as mobile technology develops in order to guarantee that consumers can continue to rely on mobile devices in their everyday lives. | http://dx.doi.org/10.12785/ijcds/150147 | Android; Cyber security; Malicious applications; Malware Detection; Mobile Malware; Mobile Application Security (Android,iOS) | Brijesh Y Sathwara (Gujarat Technological University - Graduate School of Engineering and Technology, India); Palvinder Singh Mann (Gujarat Technological University, India) |
| 48 | 1570912913 | Toward A Systematic Evaluation Approach Of Point-Of-Interest Recommendation Algorithms Of A Novel Smart Tourism Tool | Intelligent tourism can increase the interest of nomadic tourists in discovering new cities. However, many Points of Interest (POIs) are available, creating an information overload for tourists when choosing POIs to visit. For this reason, CARS (Context-Aware Recommendation Systems) can play an important role by exploiting the experiences of previous tourists and their contexts to recommend attractive POIs. Consequently, choosing the right POI recommendation algorithm (RA) for CARS is crucial because it involves the costly intervention of real tourists during the test phase. In order to make this phase more cost-effective, we can test several RAs simultaneously in order to assess their limitations in terms of cold start and tourist satisfaction. To compare these RAs, we propose in this article an approach called SEPRA (Systematic Evaluation for POI Recommendation Algorithms), which allows us to carry out an initial online evaluation of each tourist during their visit and a second offline evaluation of each CARS after the end of each POI path. To achieve this objective, we designed and implemented a new smart tourism tool that makes POI recommendations using two algorithms: the first is based on tourist/tourist similarity, and the second uses POI/POI similarity. These algorithms use memory-based collaborative filtering and are executed in parallel by our tool in the form of CARSs, incorporating time or weather as context variables. To evaluate these systems during their test phases, Our approach enables: (1) the calculation of prediction accuracy; (2) the examination of the relevance of the recommended POIs; and (3) the estimation of the acceptance rate of the recommendation process. Finally, the experimental results obtained with our approach show that the algorithm based on tourist similarity is more resistant to the cold start problem during the test phase and has a better satisfaction rate than the algorithm based on POI similarity. | http://dx.doi.org/10.12785/ijcds/150148 | Smart tourism; CARS; point of interest recommendation; collaborative filtering; online evaluation; offline evaluation | HadjHenni Mhamed (Djillali Liabes University of Sidi Bel Abbes, Algeria); Dennouni Nassim (Higher School of Management Tlemcen, Algeria); Slama Zohra (EEDIS Laboratory Djillali Liabes University Sidi Bel Abbes, Algeria) |
| 49 | 1570913259 | Predicting Apple Yield Based On Occurrence of Phenological Stage in Conjunction With Soil And Weather Parameters | Accurate and reliable yield forecasting is required for efficient planning and management of an important crop like apple. Efforts have been made to predict apple yield, mostly through the use of statistical tools with limited indicator parameters. The proposed neural network (NN) based system predicts yield of apple crops in an orchard based on identification, characterization, time of arrival and duration of phenological stages interactively with soil and weather parameters. The task of automatic yield prediction in orchards is challenging. Despite the significant amount of work that has been put into developing automated methods for estimating yields, the majority of methods currently in use are based on fruit counting, which is only useful one to four weeks before harvest. Whereas, in the proposed system, we will be predicting yield, during each phenological phase, among six classes, taking into account time of phenological stage occurrence (i.e., early occurrence, normal occurrence, or delay occurrence), soil parameter, and parameter related to weather conditions. This model will help the growers to timely take decision to execute contingency plans in case of average or low yield. The f1-score of the proposed system is 0.94. It is compared with other popular machine learning (ML) algorithms like Logistic regression, Support vector machines (SVM) and K-nearest neighbors (KNN) | http://dx.doi.org/10.12785/ijcds/150149 | Artificial Neural Network; Deep Learning; Yield prediction | Rakesh Mohan Datt (Chitkara University, India); Vinay Kukreja (Chitkara University, Punjab, India) |
| 50 | 1570918691 | School Threat Assessment System (STAS) - Recognizing Psychosocial Attributes Indicative of Violent Behavior in Students using Deep Learning | According to the Center for Homeland Defense and Security, in the first half of 2022, 2 active school shooters and 151 non-active shooter events resulted in 150 victims; the previous year's statistic the highest it has been since 1970. Most students displayed signs of mental illness and troubled behavior that was often overlooked. This research seeks to identify signs of a threat in order to distinguish and assist students who are at risk for violent behavior. 30 randomly selected shooters were analyzed through the processing of news reports to identify recurring psychosocial attributes using a WordCloud generator. A feed forward neural network then uses these traits to recognize and categorize potential growing threats in a student body. Data is collected through deep learning graphological parameters in students' handwriting using a 2D convolutional neural network. This model, with an overall accuracy of 97%, classifies cases based on the combination of 28 features that appeared in the initially studied cases. It generates an accessible report that quickly identifies students in need of immediate support, reducing the number of active-shooter incidents. The School Threat Assessment System (STAS) is available online to school systems working to increase the safety of their students from within. | http://dx.doi.org/10.12785/ijcds/150150 | Convolutional Neural Network (CNN); Feed-Forward Network (FFN); Mental Illness; School Safety | Sara Elanchezhian (Thomas Jefferson High School for Science and Technology, USA); Prommy Sultana Hossain (George Mason University, USA); Jia Uddin (Woosong University, Korea (South) & Ai and BigData, Korea (South)) |
| 51 | 1570928970 | Forecasting Students' Success to Graduate Using Predictive Analytics | Predictive analytics is the process of forecasting outcomes based on historical data. Execution of predictive analytics involves data collection, analysis and massaging, identifying machine learning, predictive modeling, predictions, and monitoring. Among phases, data analysis and massaging or data preprocessing plays a vital role in the prediction's result. This study aims to predict the student's probability of graduating on time using the student's demographic profile, previous academic achievement, and college admission result. The dataset was acquired from Caraga State University with 2207 samples of new entrants. This study implemented KNN to impute numerical data, while mode imputation was used for categorical values. Moreover, binary encoding was employed for nominal data to prevent algorithm ranked the values in order. Seven (7) algorithms were tested to the original dataset and compared to datasets integrated with LASSO (L1), Ridge (L2) regression, and Genetic Algorithm (GA) separately. The result shows that L1 with Decision Tree classifier has the lowest accuracy (58%) and AUC score (50%). It also has the smallest number of features selected (5). GA, on the other hand, selected thirty-three (33) features with AUC score of 71% and predicted 79% accurately using the Logistic Regression classifier. It exhibited 21% increase in AUC score compared to no feature selected dataset (NFS) with the same classifier. | http://dx.doi.org/10.12785/ijcds/150151 | feature selection; genetic algorithm; predictive analytics | Jayrhom R. Almonteros, Junrie B. Matias and Joanna Victoria S. Pitao (Caraga State University, Philippines) |
| 52 | 1570952039 | Real Time System Based Deep Learning for Recognizing Algerian Sign Language | Sign language plays a crucial role in facilitating communication and interaction for the deaf community. However, the recognition of sign language poses unique challenges, especially in the context of Algerian Sign Language (ALGSL), where limited research has been conducted. Using recent advances in the field of deep learning, we present a novel ALGSL recoginition system using hand cropping and hand landmarks from successive video frames. Also, we propose a new key frame selection method to find a sufficient number of successive frames for the recognition decision, in order to cope with a near real-time system, where tradeoff between accuracy and response time is crucial to avoid delayed sign recognition. Our system is based on Autoencoder architecture enhanced by attention mechanism. The Autoencoder architecture combines both convolutional neural networks (CNN) for capturing spatial information and long-short-term memory (LSTM) for capturing temporal information. The proposed architecture is evaluated on our new ALGSL dataset and achieved an accuracy of 98,99%. Additionally, we test our architecture on different publicly datasets and shows outstanding results. Finally, we present the proposed ALGSL recognition system based on Autoencoder model. | https://dx.doi.org/10.12785/ijcds/150152 | Algerian Sign Language recognition; Deep learning; Convolutional neural networks; Long-short-term memory; Attention; Mediapipe | Ahmed Kheldoun, Imene Kouar and El Bachir Kouar (University of Medea, Algeria) |
| 53 | 1570869541 | Deep Learning Based Person Authentication System using Fingerprint and Brain Wave | The practice of automatically recognizing the correct person using computational methods based on features maintained in computer systems is known as person authentication. Security, robustness, privacy, and non-forgery are the critical aspects of any person authentication system. Traditional biometric-based systems are dependent on the use of a single modality, which may be lacking in the ability to provide high security. These systems are vulnerable to noise and can be readily exploited. An optimization enabled deep learning based multimodal person authentication system is presented to solve these disadvantages. Here, a combination of brainwave signals and fingerprint images are utilized for providing improved security. A Deep Maxout Network (DMN) is utilized for performing person authentication on both the modalities and the output obtained is fused together using cosine similarity to attain the final result. The African vultures-Aquila Optimization (AVAO) algorithm is a unique optimization algorithm for updating the DMN weights. To construct the algorithm, the African Vulture Optimization Algorithm (AVOA) techniques are updated according to the extended exploration capabilities of the Aquila Optimizer (AO). The presented multimodal person authentication system achieves an accuracy of 0.926, sensitivity of 0.940, specificity of 0.928, and F1-score of 0.921, demonstrating exceptional performance. The experimental study also indicates the performance evaluation comparison of AVAO with the prevailing techniques such as Multi-task EEG-based Authentication, Multi model-based fusion, Multi-biometric system, and Visual secret sharing and super-resolution model on the basis of various metrics. | https://dx.doi.org/10.12785/ijcds/150153 | Person authentication; multimodal; fingerprint; brain signal | Rasika Deshmukh (Fergusson College (Autonomous), Pune, India); Pravin Yannawar (Dr Babasaheb Ambedkar Marathwada University, Aurangabad, India) |
| 54 | 1570912268 | Big data (BD)-based approach to network security (NS) and intelligence | The purpose of this research is to analyze the importance of network security in today's society where the internet has become essential in everyday life. With the increasing global concerns about cybersecurity, it is crucial to use big data analysis technologies and machine learning techniques to examine and forecast the state of network security. The existing models for network security monitoring have challenges such as being resource-intensive, inaccurate analysis results, low processing efficiency, and unsuitability for real-time and large-scale scenarios. To overcome these challenges, the study proposes a new BP neural network-based model that incorporates the blocked fuzzy C-means clustering approach to enhance the input data's characteristics and improve the model's accuracy. The model's methodology is comprehensively explained, and testing is conducted to verify its accuracy and usefulness in perceiving network security scenarios. The proposed model has the potential to overcome the challenges faced by current network security monitoring models and provide a viable solution for network security analysis and prediction. | https://dx.doi.org/10.12785/ijcds/150154 | Network Security, Big Data Analytics, Machine Learning, BP neural network, Fuzzy C-means Clustering, Situation Awareness, Data Preprocessing, Accuracy, Cyber Security | Mubarak Alquaifil, Shailendra Mishra and Mohammed Alshehri (Majmaah University, Saudi Arabia) |
| 55 | 1570931120 | An Intelligent Grading System for Automated Identification and Classification of Banana Fruit Diseases Using Deep Neural Network | This study proposes an intelligent system for automated illness diagnosis and categorization in banana fruit, as well as an integrated grading system. To accomplish accurate illness identification and grading, the suggested system incorporates computer vision methods, machine learning algorithms, and deep learning models. The system extracts key information from banana fruit images using image processing techniques, which are subsequently input into a trained classification model. The categorization model uses cutting-edge algorithms to categorise the banana fruit into several illness groups. Furthermore, the sophisticated grading system evaluates the severity and quality of the diseased fruit based on a variety of characteristics such as size, colour, and texture. The experimental findings reveal that the proposed method is successful, with high accuracy in illness diagnosis and accurate banana grading. This automated technology provides a time-efficient and cost-effective approach for disease control in banana plantations, allowing producers and agricultural stakeholders to make more informed decisions. | http://dx.doi.org/10.12785/ijcds/150155 | Diseases; Image processing; Deep Neural networks; Banana fruit; feature extraction and Classification | Himanshu Bhaidas Patel (KBNMU Jalgaon & D N PATEL COE SHAHADA, India); Nitin Jagannath Patil (D. N. Patel College of Engineering, India) |
| 56 | 1570934223 | Matrix Factorization and Cosine Similarity based Recommendation system for cold start Problem in e-commerce Industries | A recommendation system, also referred to as a recommender system, is a type of artificial intelligence (AI) algorithm that uses big data to suggest or recommend more products to new and existing customers. It uses data to forecast, focus, and locate what users are looking for among a constantly expanding array of possibilities. Many factors, including prior purchases, search history, demographic information, and other factors, may be used to identify them. It enables the customers to find goods and services for individuals that they unable to discover in order to assist them. At the initial level with a new customer, due to lack of knowledge, the Recommender System (RS) has a cold-start issue while making suggestions. There is no history of the choices or interactions of new users when they sign up to the system. It is impossible for the algorithm to provide tailored suggestions without this data. Additionally, when a new item is added to the system, it does not already have any interactions or preferences with other objects. This makes it difficult for the algorithm to suggest the item based on user preferences. The cold-start problem may limit the utility of RS since it makes it challenging for the system to provide trustworthy recommendations because it lacks information about people and products. Customers may not use the system again as a result of this, which could lead to a negative user experience. Recommender systems may employ a variety of strategies to circumvent the cold start problem, including content-based filtering, collaborative filtering, hybrid systems, knowledge-based systems, and demographic data. This paper proposed a Cosine Similarity (CS) and Matrix Factorization (MF) method to tackle the scarity problem which further resolves cold-start problems. | http://dx.doi.org/10.12785/ijcds/150156 | Recommendation System; Collaborative Filtering; Cold-Start; Matrix-Factorization; Cosine Similarity; Sparsity | Jameel Ahamed (Maulana Azad National Urdu University, India); Mumtaz Ahmed (Jamia Millia Islamia, India); Md Nadeem Noori (Maulana Azad National Urdu University, India) |
| 57 | 1570967983 | Exploring Progress in Forest Fire Detection, Prediction, and Behavior: An In-Depth Survey | Forest fires are a major environmental challenge that pose a threat to both human life and ecological health. To effectively prevent and manage forest fires, it is crucial to have reliable detection, prediction, and behavior analysis systems in place. This paper provides a comprehensive survey of the different approaches and techniques used for forest fire detection, prediction, and behavior analysis. It covers ground-based and aerial surveillance systems, remote sensing technologies, machine learning-based approaches, and social media-based systems. The paper also discusses the challenges and limitations of current systems and provides insights into future directions for research and development in this field. Overall, this paper highlights the importance of leveraging multiple data sources and analysis methods to improve our understanding of forest fire behavior and develop effective strategies for managing this environmental threat | http://dx.doi.org/10.12785/ijcds/150157 | Forest Fire; Ground Based System; Fire Behaviour analysis; Fire Managment; Fire Detection; Fire Prediction | Ahmad Alkhatib and Khalid Jaber (Al-Zaytoonah University of Jordan, Jordan); Hassan Al-zoubi (ZUJ, Jordan); Mohammad Abdallah (Al-Zaytoonah University of Jordan, Jordan); Mosa Salah (Al-zaytoonah University of Jordan, Jordan) |
| 58 | 1570894322 | Routing Approaches used for Electrical Vehicles Navigation: A Survey | The growing demand for Electric Vehicles (EVs) depends on the high integration of this technology in many areas. Therefore, an important area of research raises interest in finding the optimal path-planning solution for electric vehicles. This paper discusses several reviews and analyzes some of the constraints of the techniques used to improve these systems. The paper discusses common models used in Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGV). This paper investigates the planning approaches that lead to finding the optimal route for the tour from the source to the destination. The review outlines the different models and systems of Unmanned Ariel Vehicles (UAV) and Unmanned Ground Vehicles (UGV). This paper can be considered as comprehensive survey research for EV routing techniques to assist researchers choose the appropriate approach for developing a system based on optimization techniques, machine learning, or Hybrid Approaches (HAs) techniques. Optimization techniques are mostly used to find the optimal path and achieve multi-objective goals. Some findings were approved as the best models inspired by natural biological as genetic algorithms, Particle Swarm Optimization, and Ant Colony Optimization. In addition to machine learning techniques as Reinforcement Learning. The hybrid approach techniques that combine optimization and machine learning techniques can increase robustness in solving routing problems. | http://dx.doi.org/10.12785/ijcds/150158 | Electronic Vehicles (EV); Unmanned Aerial Vehicles (UAV); Unmanned Ground Vehicles (UGV); Path Planning | Shaima Ebraheem Hejres (University Of Bahrain & Ministry Of Education, Bahrain); Amine Mahjoub (University of Bahrain, Bahrain); Nabil Hewahi (UOB, Bahrain) |
| 59 | 1570930557 | Food Delivery Service Applications in Highly Urbanized Cities: A Scoping Review | In recent years, food delivery services utilizing apps and websites have gained immense popularity worldwide due to the convenience it provide in highly urbanized cities across various countries. This service involves delivering food from restaurants, fast-food chains, local food joints and groceries stores using mobile applications or websites to place orders. However, there are still issues with food delivery services, such as late deliveries, poor quality food, and problems with payment and delivery drivers. The objective of this study was to conduct a comprehensive and exploratory review of the literature related to food delivery service applications (FDSA) and eventually identify possible research topics to meet the research gap for future studies. It gathers and analyzes relevant information and insights from the collected literature and provides an overview of the existing reviews on content-context-process of food delivery service applications. There were 174 relevant empirical studies on food delivery service applications identified through Google Scholar as the primary search engine in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 and the Content, Context, and Process (CCP) Framework. Through Google Scholar, 174 empirical studies were identified, with 30 eligible journals undergoing rigorous analysis. The results identified 5 content-related, 2 context-related, and 4 process-related aspects relevant to food delivery service applications. The scoping reviews of these literatures recommends for future studies on features-functionalities-designs of the FDSA, efficiency-effectivity-productivity of FDSA services and exploration of FDSA models to optimize its processes. | http://dx.doi.org/10.12785/ijcds/150159 | Food Delivery System Applications; PRISMA; Content Analysis | Clark James Semblante (Mindanao State University-Iligan Institute of Technology, Philippines); Paul B Bokingkito, Jr. (Mindanao State University - Iligan Institute of Technology, Philippines); Rogena Catanduanes (Mindanao State University-Iligan Institute of Technology, Philippines); Adrian B. Martin (Zamboanga Peninsula Polytechnic State University, Philippines); King Jehu II Radaza (Central Mindanao University, Maramag, Bukidnon, Philippines); Lemuel Clark Velasco (Mindanao State University-Iligan Institute of Technology & Premiere Research Institute of Science and Mathematics, Philippines) |
| 60 | 1570958000 | Hybrid Deep Learning Approach for Classification and Analysis of X Posts on Russia Ukraine War | X (formerly Twitter) has become a vital information source on a variety of political, social, and economic concerns, as a consequence of its growth and popularity has resulted in an enormous number of people sharing their opinions on a wide range of topics. To determine people's emotions about the Russia-Ukraine war (RUW), this study examines trends in English-language tweets. In this work, we have engaged 34 countries to tweet opinions that produce a strong perception of the people about the war and message to the world what people famine from the countries and that affects their lives. To analyze positive and negative emotions in tweets, which are represented by hope and fear, the LSTM-CNN model is based on deep learning. A time series is calculated that correlates with the frequency of positive and negative tweets in different nations. Additionally, an approach based on the neighborhood average has been used for modelling and grouping the time series of various countries. In the results, the clustering method gives results as significant information, how people feel about this dispute and share their opinions about RUW is approached. This research study helps the uninfluenced press members to have an impartial source of information for their reports and articles. | http://dx.doi.org/10.12785/ijcds/150160 | Clustering; CNN; LSTM; Social Network Mining | Amrit Suman (Sharda University, Greater Noida, India); Sudeep Varshney and Kuldeep Chouhan (Sharda University, India); Preetam Suman (VIT Bhopal); Gunjan Varshney (UTU, India) |
| 61 | 1570948906 | A Systematic Literature Review of an Advisory System | The growth of artificial intelligence (AI) driven technologies has been proposed as a means to improve the standard of people's lives. The advent of advisory system has manifested as significance element in artificial intelligence that effectively helping people in various field. This study initially presents 453 articles by examining the literature between 2015 and 2022. After a meticulous review process, the studies were filtered down to 59 articles for full analysis. This review provides significant contributions in finding into the exploration of advisory system, specifically in the existing framework of advisory system, technique applied in advisory system, domain in which the artificial intelligence technique is being applied in the advisory system and the validation technique use in validating the advisory system. | http://dx.doi.org/10.12785/ijcds/150161 | Advisory System; Advisory System Framework; Artificial Intelligence; Systematic Literature Review | Ain Nadhira Mohd Taib (University Malaysia Pahang Al-Sultan Abdullah, Malaysia); Fauziah Zainuddin (Universiti Malaysia Pahang, Malaysia); Rahmah Mokhtar (FSKKP & Universiy Malaysia Pahang, Malaysia); Kamalia Azma Kamaruddin (Universiti Teknologi MARA, Malaysia); Nadzirah Ahmad Basri (International Islamic University Malaysia, Malaysia) |
| 62 | 1570932547 | Real Time Oral Cavity Detection Leading to Oral Cancer using CNN | Oral cancer poses a substantial global health threat, as it continues to witness escalating incidence rates and consequential mortality on a widespread scale. To enhance patient outcomes, the crucial role of early detection cannot be overlooked. This research introduces an innovative real-time approach to detect various oral cavity conditions, focusing specifically on the prediction of oral cancer using a deep learning framework. Our methodology integrates patient questionnaires and oral cavity images, amalgamating them to improve the accuracy and reliability of our predictive model. The comprehensive questionnaires gather extensive data on dietary habits, lifestyle factors, and potential risk factors associated with oral cancer. Leveraging deep learning models such as ResNet101, ResNet50, ResNet152, and VGG19, we classify oral cavity images as either cancerous or non-cancerous. By considering the relative weightage of the questionnaire responses and image analysis predictions, we compute a final probability of oral cancer. A diverse dataset is utilized to evaluate the performance of our proposed model, assessing its accuracy, sensitivity, specificity, and overall predictive capability. The resulting system aims to provide healthcare professionals with a real-time prediction tool featuring a user-friendly interface, thereby facilitating early detection and intervention. The outcomes of this study significantly contribute to the advancement of oral cancer detection methods, offering the potential to enhance patient outcomes through timely intervention. | http://dx.doi.org/10.12785/ijcds/150162 | Convolutional Neural Networks; Deep Learning; Oral Cancer; Oral Potentially Malignant Disorder | Shruthi K (Siddaganga Institute of Technology & SIT, India); poornima a. s. Kulkarni, Musaddiq Shariff, Pradeep Singh S M, Subramanyam Subramanyam D P and Varun M H Varun (Siddaganga Institute of Technology, India) |
| 63 | 1570942729 | Detection Tampering in Digital Video in Frequency Domain using DCT with Halftone | This In recent years, the rapid technological development and the emergence of mobile devices, cameras, etc., in addition to the availability of video production, editing, and formatting programs, made it easy to edit, manipulate, and fake or tamper video. As they know that pictures or videos give more information than texts; Video is a very important medium for transferring information from one place to another. One of the important types of evidence in road accidents and theft crimes. Moreover, when forensic analysis is required for any video, the source video is not available and the forensic experts must make decisions based on the existing video (under surveillance) and determine if this video is fake (tampered) or not fake. There are multiple methods to tamper video, including active and blind passive methods. In this research, we tried to combine the behavior of active methods in the process of embedding the halftone current frame of video in the DCT Coefficients of next frame of the same video with the behavior of passive methods by comparing the information embedded after extracting with the information of the current frame to determine whether there is a fake in the video or not & which frame contains tamper. The experimental results of the proposed method showed a high level of success in locating frames in which falsification or tampering occurred through copying, deletion or insertion, or even if copy-move regions. Also, in proposed method we attempted to post-processing the fake frames using the information included in the subsequent frame, if it is not faked. Finally, the original video, the embedded halftone video, and the tamper (fake) video after post processing were compared using PSNR and SSIM similarity scales. At last, the accuracy & precision scores of tampered & non tampered frames are computed. | http://dx.doi.org/10.12785/ijcds/150163 | Video tampering; DCT transform; Halftone algorithm; PSNR- SSIM Similarity measures; Gaussian filter | Wafaa Hasan Alwan (University of Kerbala, Iraq); Sabah M Alturfi (University of Karbala, Iraq) |
| 64 | 1570954631 | A Dynamic Indoor Localization with Movement Validation and Fingerprinting Technique under IEEE 802.15.4 Network | The study of indoor localization has been extensively studied, either in terms of wireless technologies or localization techniques. The accuracy is then challenged when the monitored object is actively moving. Previously, we studied a continuous and low-power fingerprint-based indoor localization system using IEEE 802.15.4 (FILS15.4), which has been integrated into a smart environmental IoT platform. Although fingerprint-based localization offers a great advantage in its simplicity, it relies on real-time signal strength measurements and databases. Thus, it suffers challenges in accuracy when the object is continuously moving. In this study, we focus on developing dynamic positioning, where users continuously move from one room to another. Due to human movements, the fluctuation of the link quality indicator (LQI) can affect the detection accuracy. To avoid false detection, we propose a movement validation method by checking the variance of the LQI and accelerometer to differentiate the cause of fluctuations and increase the detection accuracy. For experiments, we run the test-bed of {\em FILS15.4} on a two-floor layout. Five to six receivers were allocated to detect multiple users. The results show that the system yields 96.2% accuracy using six receivers simultaneously. Thus, it gives sufficient detection accuracy even for dynamic conditions. | http://dx.doi.org/10.12785/ijcds/150164 | indoor positioning; accelerometer; motion sensing; IoT; energy efficient | Pradini Puspitaningayu (Universitas Negeri Surabaya, Indonesia); Nobuo Funabiki, Yuan-Zhi Huo and Yohanes Yohanie Fridelin Panduman (Okayama University, Japan) |
| 65 | 1570958381 | Ransomware Detection Dynamics: Insights and Implications | This study provides a multifaceted analysis of ransomware-related data, offering insights that underscore the complexity and evolving nature of cybersecurity threats. Our exploration of financial aspects revealed the absence of a fixed ransom amount associated with specific ransomware types, highlighting the adaptability of threat actors. Moreover, the correlation analysis unveiled a strong link between ransomware clusters and cryptocurrency transaction patterns, enhancing the potential for predictive and preventive cybersecurity measures. The temporal analysis emphasized critical time intervals during ransomware attacks, guiding the development of timely response strategies. Collectively, these findings emphasize the importance of data-driven, adaptive cybersecurity approaches to effectively address the ever-changing landscape of ransomware threats, safeguarding organizations and individuals against potential cyberattacks. | http://dx.doi.org/10.12785/ijcds/150165 | Cybersecurity; Ransomware Detection; Machine Learning; Feature Selection; UGRansome Dataset | Mike N Wa Nkongolo (University of Pretoria, South Africa) |
| 66 | 1570931849 | A Cooperative Relaxation-Based Method for Range Image Segmentation | In this work, we propose a combination of two methods for the segmentation of range images (Depth Image), based on the relaxation of the detection results of both edges and regions. The latter are obtained from two different representations of the data. The principle of the introduced method consists in matching the two detection results as a mutual self-regularization, and whose objective is to produce an optimal final segmentation with respect to a given criterion of optimality, expressed by a well appropriate objective function. The experimental results, using the ABW database, show the potential of the proposed method for efficient segmentation of range images. | http://dx.doi.org/10.12785/ijcds/150166 | Range Image; Edge detection; Region-based segmentation; Relaxation | Imene Belloum and Mourad Bouzenada (University Abdelhamid Mehri, Constantine2, Algeria); Smaine Mazouzi (Université 20 Août 1955 - Skikda, Algeria) |
| 67 | 1570946787 | Improving Detection and Prediction of Traffic Congestion in VANETs: An Examination of Machine Learning Algorithms | Traffic congestion remains a pressing challenge in urban areas, causing significant economic and environmental repercussions. To address this issue, accurate detection and prediction of traffic congestion are imperative for effective traffic management and planning. This research study investigates the efficacy of Support Vector Machines (SVM) and various other machine learning algorithms in augmenting traffic congestion detection and prediction for Vehicular Ad hoc Networks (VANETs). Leveraging historical congestion patterns, we train and evaluate the performance of the algorithms. Our results demonstrate the potential of SVM, coupled with advanced feature engineering techniques, to outperform other methods in accurately identifying and forecasting traffic congestion. The SVM classifier achieved an impressive classification accuracy of 0.99, showcasing its effectiveness in handling diverse traffic scenarios. Additionally, the K-Nearest Neighbors (KNN) and Ensemble Learning classifiers also yielded commendable accuracies of 0.99. Notably, the Decision Tree (DT) classifier attained a perfect accuracy score of 1.00, indicating its robustness in handling congestion patterns. The proposed approach not only achieves high detection accuracy but also exhibits remarkable robustness and scalability, enabling its application across various traffic scenarios. These findings contribute significantly to the development of intelligent traffic management systems, providing valuable insights into optimizing transportation networks. Ultimately, implementing our approach holds the potential to alleviate congestion, enhance travel efficiency, and foster urban sustainability | http://dx.doi.org/10.12785/ijcds/150167 | Traffic congestion; Support Vector Machines (SVM); Ensemble Learning | Mohammed S Jasim (Safx University, Tunisia); Nizar Zaghden (SETIT, Tunisia); M. Bouhlel (University of Sfax, Tunisia) |
| 68 | 1570939259 | Computer-Assisted Disease Diagnosis Application for Malaria Early Diagnosis Based on Modified CNN Algorithm | Since its emergence in the early 20th century, Malaria has been confirmed as a deadly disease that has spread throughout the world with very high mortality and morbidity. This is in accordance with the WHO report in 2018 which stated that worldwide there have been more than 220 million cases of malaria with a death rate of nearly 500 thousand cases. However, Malaria is actually a disease that can be cured and prevented if treatment initiatives are implemented early and effectively. Unfortunately, this disease is often ignored because it is considered the common cold and is only diagnosed when it has reached a critical phase. This research is expected to be an alternative for early diagnosis of malaria. Hence, confirming the presence of the malaria parasite earlier will make the treatment of this disease more effective in reducing mortality. This research is expected to produce website-based computer-assisted disease diagnosis (CAD) software enriched with deep learning algorithms to become an alternative for early diagnosis of malaria. This CAD system has the potential to provide fast and reliable malaria diagnosis and avoid detection errors by experts due to human error. In addition, this study applies a modified Pre-trained Convolutional Neural Network (CNN) architecture to improve the performance of the developed deep learning model. Based on the results obtained, this study succeeded in exceeding the benchmarks of previous studies with an accuracy value above 97%. This CAD software is also equipped with various features to make it easier to use. | http://dx.doi.org/10.12785/ijcds/150168 | Computer-assisted disease diagnosis (CAD); Convolutional Neural Network (CNN); Malaria; Pre-trained CNN; Thin Blood Images; Website | Zul Indra (Universitas Riau, Indonesia); Yessi Jusman (Universitas Muhammadiyah Yogyakarta, Indonesia); Elfizar Elfizar (University of Riau, Indonesia); Roni Salambue (Universitas Riau, Indonesia); Rahmad Kurniawan (Universitas Riau, Indonesia & Universiti Kebangsaan Malaysia, Malaysia); Tisha Melia (Universitas Riau, Indonesia) |
| 69 | 1570945880 | The Use of Motion Capture Technology in 3D Animation | Main Objective: This objective of this study is to provide a thorough overview of the process and development possibilities of motion capture (Mo-Cap), focusing on the relationship between accessibility, simplicity, and data quality. Background Problem: 3D modeling is popular for IT-based applications because of its reusability, detailed breakdown, and ability to depict depth. Motion capture is growing in popularity in animation production, because of its advantages like interactive features, lifelike representation, and automated functionality. Nonetheless, the fusion of motion capture and 3D animation presents intricate obstacles and developmental hurdles. Novelty: This research explores the application of Mo-Cap technology in the realm of 3D animation besides that the potential of algorithms in understanding human motion and improving motion capture technology is also discussed. Through a comprehensive review, this study identified the advantages and limitations of motion capture systems and the importance of balancing accessibility and quality in the process. Research Method: This study is a literature review that aims to identify challenges and opportunities in the application of motion capture for 3D animation. Finding/Result: This research provides an overview of motion capture in 3D animation, including the possibilities and challenges. The study also analyzes the advantages and disadvantages of different motion capture techniques systems and highlights the importance of simplifying the system without sacrificing quality. The results of this study emphasize the importance of considering real-time quality and usability when selecting a method for a particular application. Conclusion: Motion capture has a broad spectrum of uses in the field of 3D animation from high-cost film production to products targeted towards consumers. This study emphasizes the importance of balancing accessibility and quality in selecting motion capture systems and methods. In addition, this research reveals the potential of machine learning algorithms in driving the development of motion capture technology. In summary, this study offers an extensive examination of the motion capture procedure, as well as its potential and obstacles within the domain of 3D animation | http://dx.doi.org/10.12785/ijcds/150169 | motion capture; 3D-Model; 3D-Animation | Mars Caroline Wibowo (STEKOM University, Indonesia); Sarwo Nugroho (STEKOM University Semarang Indonesia, Indonesia); Agus Wibowo (University of Science and Computer Technology, Indonesia) |
| 70 | 1570946234 | ArabAlg: A new Dataset for Arabic Speech Commands Recognition for Machine Learning Purposes | Automatic Speech Recognition (ASR) systems have witnessed significant advancements in recent years due to the emergence of deep learning techniques and the availability of large speech datasets. With the increasing demand for Arabic voice-enabled technologies, the availability of a high-quality and representative data set for the Arabic language becomes crucial. This paper presents the development of a new dataset called \textbf{ ArabAlg} specifically designed for Arabic Speech Commands Recognition (ASCR) to support the integration of Arabic voice recognition systems in smart devices in Internet of Things (IoT). This research focuses on collecting and annotating a diverse range of Arabic speech commands, encompassing various domains and applications. The dataset construction process involves recording and preprocessing several utterances from native Arabic speakers. To ensure precision and reliability, quality control measures are implemented during data collection and annotation. The resulting dataset provides a valuable resource for training and evaluating ASCR systems tailored for Arabic speakers using Machine Learning and Deep Learning. | http://dx.doi.org/10.12785/ijcds/150170 | Automatic Speech Recognition; Machine learning; Dataset for limited vocabulary; Arabic Speech Commands Recognition; Smart devices; Internet of Things | Nourredine Oukas (University of Bouira, Algeria & University of Boumerdes, Algeria); Samia Haboussi (University of Bouira, Algeria); Chafik Maiza and Nassim BenSlimane (Akli Mohand Oulhadj University of Bouira, Algeria) |
| 71 | 1570907376 | Dynamic Traffic Management using Reinforcement Learning | Traffic congestion has become a major problem in this rapidly growing world. Everyone operating a vehicle, as well as the traffic police in charge of managing the traffic, finds it difficult to become stuck in heavy traffic. For this a set, predetermined timing for traffic flow for each direction at the junction is utilized by traditional traffic light controllers. However, the concept of a fixed time traffic signal controller does not work well in places with chaotic traffic patterns. A dynamic traffic control system is therefore required, which regulates the traffic signals in accordance with the volume of traffic. This paper proposes a model that uses reinforcement learning (RL) along with deep neural networks (DNN) to manage discretions (signal status) for the proffered environment with the help of Simulation of Urban MObility (SUMO). The main objective of this research study is to construct a model that can independently determine the best course of action and aims to provide better traffic management that will decrease the average waiting time, cause lower congestion, and provide a smooth flow of traffic. | http://dx.doi.org/10.12785/ijcds/150171 | Dynamic Traffic Management; Traffic Patterns; Reinforcement Learning; Simulation of Urban MObility; Deep Neural Networks; Deep Q-learning | Aryaan Shaikh (Vishwakarma Institute of Information Technology, Pune, India); Babasaheb Bhalekar and Pravin Futane (Vishwakarma Institute of Information Technology, India) |
| 72 | 1570867135 | Multi-Voltage Design of RISC Processor for Low Power Application: A Survey | Power management is becoming important aspect as the size of transistor is shrinking. For processor design, Reduced Instruction Set Computer (RISC) architecture is preferable as compared to Complex Instruction Set Computer (CISC) architecture because of its simplicity and availability. To design the low power RISC processor, there are a few techniques that had been used earlier, such as a) pipelining and b) Common Power Format language to generate power intent of RISC processor design. In the present work, for designing a 16-bit RISC processor with low power consumption, a multi-voltage design technique has been used. In this technique, different supply voltages are provided to different blocks of the design. This technique is implemented with the help of Unified Power Format (UPF). Further, various operations such as ADD, SUB, INVERT, AND, OR, Right Shift, Left Shift, and Less Than are verified on modelsim for the designed 16-bit RISC processor. | http://dx.doi.org/10.12785/ijcds/150172 | Unified Power Format (UPF); Low Power Design; Multi-Voltage Design; Power Gating; Clock Gating; Retention | Dheeraj Kumar Sharma (National Institute of Technology, Kurukshetra, India); Rahul Vikram (Qualcomm Inc India, India) |
| 73 | 1570908321 | Forensics Analysis of Cloud-Computing Traffics | Forensics analysis of cloud-computing traffic involves investigating and analyzing network traffic in a cloud-computing environment to identify any suspicious or malicious activity. This can include identifying and analyzing network connections, identifying and analyzing data transfer patterns, and identifying and analyzing data payloads. It's also necessary to have a good understanding of cloud infrastructure and network protocols to effectively analyze cloud-computing IoT traffic. The increasing use of IoT device traffic in cloud computing environments has led to a need for effective forensic analysis methods in order to investigate and respond to security incidents. This thesis examines the challenges and considerations for forensic analysis in IoT-based cloud computing systems, including the unique characteristics of IoT devices and the distributed nature of cloud computing environments. We propose a framework for forensic analysis in IoT-based cloud computing systems that includes guidelines for data collection, preservation, and analysis. We also present case studies that demonstrate the effectiveness of the proposed framework in a real-world IoT-based cloud computing scenario. The results of this thesis contribute to a better understanding of the forensic challenges in IoTbased cloud computing systems and provide practical solutions to address them. | https://dx.doi.org/10.12785/ijcds/150173 | Cloud computing; IoT; Digital forensic; Data gathering; Data analysis | Moayad Almutairi, Shailendra Mishra and Mohammed Alshehri (Majmaah University, Saudi Arabia) |
| 74 | 1570915212 | Detection of Roman Urdu fraud or spam SMS in Pakistan using machine learning | Over the past few years, mobile devices and their services have become widely used around the world. Almost everyone uses the Text Messaging Service (SMS) for communication purposes because it is easy to use and inexpensive. When a person tries to deceive another for the sake of profit (material or money), it is known as Fraud. Through SMS fraud, fraudsters often adopt various strategies to make their messages look credible and legitimate. Various popular organizations use SMS services to advertise their products and send messages to individuals about their services. As a result, one receives many junk messages. Spam message is a message sent to any user who does not want to have it on their phone. Spam or fraudulent messages can be threatening and can sometimes cause financial and confidential data loss. In Pakistan, messages are sent in English and Urdu (Pakistani national language) but most messages are sent using Roman Urdu (Urdu written using Latin / English characters). This research compares the strategies and algorithms used in the literature to detect spam / fraudulent messages written in English or in any local language such as Roman Urdu. The study also suggests a new way to detect fraudulent messages written directly in Roman Urdu. In the fraud detection process, three different monitoring machine learning classifiers are used in this study namely Support Vector Machine (SVM), Naïve Bayes (NB) and Decision Tree (J48). After using the model, we found that SVM performed better than the other two classifiers with 99.42% accuracy. | http://dx.doi.org/10.12785/ijcds/150174 | spam SMS; fraud; Roman Urdu; machine learning; Detection | Muhammad Ayaz, Sarwat Nizamani, Aftab Ahmed Chandio and Kirshan Luhana (University of Sindh, Pakistan) |
| 75 | 1570910238 | Design and Implementation of a Heterogeneous Distributed Database System for Students Absence Automation Using Portable Devices | Student absence declaration is one of the basic processes that occur on a daily basis in faculties, often relying on paper forms, which is inefficient when the faculty administration wants to gather absences from multiple departments (sites) and calculate the absence rates for the faculty. In response to these challenges, a heterogeneous distributed database system was designed for a faculty to declare and automate student absences using smartphones or laptops, as well as send absences in real time from inside lecture halls or laboratories in departments to the faculty administration to generate absence reports. The system employed a client-server model to deliver services through the local area network (LAN) and the internet to the various departments and the college administration. MSSQL database tables were designed and fragmented across faculty sites using horizontal fragmentation technology, then encrypted with the AES265 algorithm. The proposed system also involved the use of private SQLite or MSSQL databases for the lecture's Android or Windows devices, respectively. Faculty administration staff, on the other hand, used a secure website built with ASP.NET to access the MySQL database server, which contains the absence table that was populated from all sites. The system was protected against false absence declarations by utilizing the local server's timing factor, which allowed the lecturer to only declare absences on specific days and times. The experimental results showed high reliability and accuracy of the system and high satisfaction among the users about achieving the desired goals. The findings demonstrated a strong preference for the system among the lecturers over other systems that relied on student devices, biometric devices, and AI-based ones, while the faculty's administration staff preferred automatic absence collection by the distributed design of the system over manual collection. | http://dx.doi.org/10.12785/ijcds/150175 | Distributed Database; Heterogeneous Database; Smartphones; Portable Devices | Ahmed A Alsamman (College of Computer Science and Mathematics, University of Mosul) |
| 76 | 1570910099 | PrinciResnet Brain Tumor Classification Technique for Multimodal Input-level Fusion Network | Brain tumors are a leading cause of mortality in India, with over 28,000 cases reported annually, resulting in more than 24,000 deaths per year, according to the International Association of Cancer Registries. Early detection, segmentation, and accurate classification are crucial in effective tumor analysis, and various algorithms have been developed to achieve this. In this study, we propose a novel approach for the detection and classification of brain tumors using both single slices of MRI and CT, as well as input-level fused images of MRI & CT. Our approach involves the implementation of the PrinciResnet brain tumor classification technique, which is based on Principal Component Analysis (PCA) and Resnet techniques. We report that our approach significantly improves the accuracy, sensitivity, and specificity parameters to 90%, 96%, and 95%, respectively, based on a dataset of 600 fused slices and 1000 single slices obtained from reputable sources. Our findings hold promise for improving the diagnosis and treatment of brain tumors, which are a significant cause of mortality globally. | http://dx.doi.org/10.12785/ijcds/150176 | Resnet; PCA; Skip connection; Brain tumor; CT; MRI | Kannan G (B S Abdur Rahman Crescent Institute of Science & Technology, India); Padma Usha Manikumar, Sai Akshay S, Giri S and Shaik Mohammed Huzaifa (B S Abdur Rahman Crescent Institute of Science and Technology, India) |
| 77 | 1570945424 | applying hybrid clustering with evaluation by AUC classification metrics | Traditional metrics may not adequately assess performance in certain situations, whereas Area Under the Curve (AUC) offers a comprehensive perspective by considering sensitivity and specificity. This method enhances interpretability, addresses limitations, and promotes the development of robust clustering algorithms. Incorporating AUC into unsupervised learning is crucial for refining data analysis strategies. This paper highlights the advantages of using AUC classification metrics as an evaluation tool for clustering algorithms inspired by a recent novel contribution utilizing the AUCC technique. Hybrid clustering models merge the strengths of various clustering approaches to deliver more robust, accurate, and adaptable solutions. Since linkages significantly influence cluster structure and cohesion, selecting the appropriate linkage is essential for accurately discerning patterns in complex datasets. Consequently, we employed single and average linkage methods using Euclidean and Manhattan distance measures. The experiment was conducted on the NSL-KDD dataset for intrusion detection purposes. The results indicated variance in False Alarm Rate (FAR) and Detection Rate (DR) across different NSL-KDD training and testing subsets. | http://dx.doi.org/10.12785/ijcds/150177 | Hybrid, classification, Clustering, Evaluation Metrics, AUC, Linkages, Distance | Ali F. Dakhil (University of Thi Qar, Iraq); Waffaa M. Ali (University of Shatrah, Iraq); Mustafa Asaad Hasan (University of Thi-Qar, Iraq) |
| 78 | 1570916355 | Open research issues and tools for visualization and big data analytics | The new age of digital growth has marked all fields. This technological evolution has impacted data flows which have witnessed a rapid expansion over the last decade which makes the data traditional processing unable to catch up with the rapid flow of massive data. In this context, the implementation of a big data analytics system becomes crucial to make big data more relevant and valuable. Therefore, with these new opportunities appear new issues of processing very high data volumes requiring companies to look for big data-specialized solutions. These solutions are based on techniques to process these masses of information to facilitate decision-making. Among these solutions, we find data visualization which makes big data more intelligible allowing accurate illustrations that have become accessible to all. This paper examines the big data visualization project based on its characteristics, benefits, issues, and challenges. The project, also, resulted in the provision of tools surging for beginners as well as well as experienced users. | http://dx.doi.org/10.12785/ijcds/150178 | Big Data; DataViz; Cloud; Big Data Analytics; Business Intelligence | Rania Mkhinini Gahar (OASIS Research Laboratory, Tunisia); Olfa Arfaoui (RISC Laboratory, Tunisia); Minyar Sassi Hidri (Imam Abdulrahman Bin Faisal University, Saudi Arabia) |
| 79 | 1570954343 | Optimizing Fog Computing Efficiency: Exploring the Role of Heterogeneity in Resource Allocation and Task Scheduling | Fog computing has emerged as a promising approach to address the stringent requirements of latency-sensitive applications in the Internet of Things (IoT) era. This research conducts a comprehensive and rigorous investigation into the performance characteristics of fog computing systems, with a specific focus on analyzing the impact of architectural heterogeneity. The study delves into the influence of architectural diversity, deadline constraints, and the count of Mobile Data Centers (MDCs) on key performance metrics. Employing extensive empirical analysis and meticulous experimental simulations, the research evaluates success ratios, rejection ratios, and resource utilization across diverse architectural models. The results underscore the paramount importance of adopting heterogeneous architectures and wider deadline ranges, resulting in significantly improved success ratios and reduced job rejections. Additionally, the study reveals the positive implications of increasing the number of MDCs on resource utilization and overall system performance. This research provides invaluable and actionable insights into optimizing fog computing systems, facilitating the development of highly efficient and effective solutions for real-world applications. | http://dx.doi.org/10.12785/ijcds/150179 | IoT; fog computing; scheduling; resource utilization | Nikita Sehgal (Giani Zail Singh Campus College of Engineering and Technology, Bathinda, India); Savina Bansal and Rakesh Kumar Bansal (GZSCCET, India & Maharaja Ranjit Singh Punjab Technical University Bathinda, India) |
| 80 | 1570939711 | IoMT-based Heart Rate Variability Analysis with Passive FBG Sensors for Improved Health Monitoring | The use of smart healthcare systems to monitor cardiac parameters has gained widespread popularity globally due to advancements in technology. The Internet of Medical Things (IoMT) has become an integral part of modern healthcare by facilitating the efficient monitoring of vital signs through advanced sensors. Heart rate variability (HRV) parameters, which provide valuable insights into a patient's condition, are now a crucial aspect of healthcare applications. Among the innovative solutions available, fiber Bragg grating (FBG) based optical sensors have emerged as a promising technology for continuous monitoring of various cardiac parameters. Recent technological breakthroughs have made these sensors highly accurate, enabling early detection and prediction of cardiac diseases, significantly impacting lives. This article focuses on the design, construction, and structural analysis of a passive optical FBG sensor capable of acquiring real-time HRV parameters such as heart rate, standard deviation of normal-to-normal intervals, root mean square of successive differences, and percentage of successive NN intervals differing by more than 50 ms. Additionally, advanced signal processing algorithms and an IoT-based architectural design are presented. An experimental study conducted in a laboratory involved five subjects, three males and two females, and demonstrated satisfactory performance with an error rate of less than 10% compared to a standard HR monitor. This intelligent system can detect arrhythmia, coronary heart disease, aortic diseases, and strokes, thereby making a significant contribution to healthcare. The combination of FBG sensors, IoT architecture, and advanced technology holds immense potential for enhancing cardiac monitoring and improving patient outcomes. | http://dx.doi.org/10.12785/ijcds/150180 | Fiber Bragg grating sensor; IoMT; SDNN; RMSSD; PNN50; PDMS | Maitri Mohanty (GIET University, Gunupur, India); Premansu Sekhara Rath (GIET University, India); Ambarish Gajendra Mohapatra (Silicon Institute of Technology, India) |
| 81 | 1570912418 | On the Impact of Radio Hub Subnet Clustering for 100 cores Mesh Wireless NoC Architecture | The integration of numerous embedded cores onto a single die is made possible by the Network-on-chip (NoC), which serves as a crucial technology. The utilization of planar metal interconnects for implementing a NoC through existing methods is inadequate due to the considerable power consumption and high latency caused by the usage of multihop channels in data interchange. Therefore, to address scalability challenges that could impact on-chip communication systems in future manycore architectures have led to the proposal of the wireless Network-on-Chip or WiNoC design paradigm as a potential solution. The impact of radio hub subnet clustering on the 100 cores mesh Wireless NoC architecture is analyzed in this paper. This study investigates the overall delay in data transmission, network throughput, and energy consumption in the 100 cores mesh Wireless NoC architecture with 4 radio hubs by analyzing single, four, and nine-tile radio subnet clustering. The validation of the results involves simulating the 100-core mesh WiNoC architectures being tested using the cycle-accurate Noxim simulator, with both random and transpose traffic workloads. According to the simulation results, the architecture with a four-tile radio hub subnet clustering provides the most optimal performance for both random and transpose traffic distributions at PIR 0.013 flit/cycle/tile when compared to the other analyzed subnet clusterings. | http://dx.doi.org/10.12785/ijcds/150181 | Network-on-Chip; Wireless; Mesh Topology; Radio Hub; Subnet Clustering | Asrani Lit (Universiti Malaysia Sarawak (UNIMAS), Malaysia) |
| 82 | 1570927026 | Studying Vibratory Patterns of Vocal Folds and their Impairments in Parkinson's Disease: A Theoretical Approach | Current vocal fold models emphasize either time or space during voice production but leave the source unconsidered. In the present review paper, we studied the vocal fold oscillations by dividing them into zone 1 and zone 2. Zone 1 includes the brain, neurons, and vagus nerve, and zone 2 comprises mechanical apparatus including lungs, voice box, and articulators. Then, a zonal view of changes in the voice, and vocal impairments due to Parkinson's disorder is demonstrated. The paper further examines various mathematical models analyzing vocal fold oscillations and their respective limitations. Since voice production is a cause-and-effect relationship, thus we used double point Green's Function method to solve the problem of vocal fold oscillation. This adds the causality factor and delay term to the problem and also illustrates the potential problems that could arise during the creation, conversion, or transition of impulses during speech production. Hence, our approach addresses the problem at the time when sound waves are in the form of electromagnetic impulses in the brain. Due to a lack of test subjects, however, we are unable to evaluate the experimental outcomes of this approach. Finally, the paper also mentions some unexplored limitations, which can be elucidated in the future. | http://dx.doi.org/10.12785/ijcds/150182 | Green's Function; Vocal Fold Vibrations; Acoustic models; Parkinson's Disease | Richa Indu and Sushil Chandra Chandra Dimri (Graphic Era Deemed to be University, India) |
| 83 | 1570923741 | Using Cloud Services to Improve Weather Forecasting Based on Weather Big Data Scraped From Web Sources | Big Data (BD) scraping systems are among the recommended approaches for large-scale web data extraction. However, these systems for collecting large amounts of data face many challenges, including processing, storage, and data extraction reliability. Due to its potentials, cloud computing is becoming a viable solution to support BD scraping systems. This paper tenders a cloud based-web scraping framework for weather BD extraction and analysis. The aim is to extract weather BD from web sources, analyze this data and use it for visualization and forecasting purposes, and this by enabling elastic and on-demand resources. The framework is implemented using Selenium and Amazon Web Services and tested with Morocco weather data. The suggested cloud-based scrapper's performance and scalability analysis reveals that it provides more efficiency in terms of data collecting and analysis, as well as forecast quality, due to its capacity to leverage cloud resources. | http://dx.doi.org/10.12785/ijcds/150183 | Cloud computing; Weather Big Data; Weather forecasting; Web scraping | Abderrahim El Mhouti, Mohamed Fahim, Asmae Bahbah, Yassine El Borji, Adil Soufi and Mohamed Erradi (Abdelmalek Essaadi University, Morocco) |
| 84 | 1570954496 | A Comprehensive Comparative Study of Machine Learning Algorithms for Water Potability Classification | Water quality prediction is of utmost importance due to the scarcity of clean water resources. Machine learning (ML) techniques, including neural networks, have gained significant importance in this field. In this study, seven ML algorithms, namely Bagging classifier, Multilayer Perceptron (MLP), Logistic regression (LR), J48, Random Forest (RF), IBk, and AdaBoostM1, were employed to assess water quality. Assessment metrics such as recall, precision, F-measure, true positive (TP) rate, false positive (FP) rate, receiver operating characteristic (ROC) area, and precision-recall curve (PRC) area were utilized to identify the most accurate model. Results revealed the superior performance of RF and MLP in multiple metrics. For RF, the precision was 0.657, recall was 0.665, and F-measure was 0.636, while for MLP, the precision was 0.654, recall was 0.663, and F-measure was 0.635. Additionally, the FP rate for RF was 0.452, and for MLP, it was 0.453. These findings highlight the effectiveness of RF, MLP, and neural networks in water quality prediction. The study contributes valuable insights for implementing accurate prediction models, supporting the sustainable management of water resources. | http://dx.doi.org/10.12785/ijcds/150184 | Artificial Intelligence; Neural Network; Water Potability; Machine Learning; Civil Engineering; Environmental Engineering | Fuad Ahmad Musleh (University of Bahrain, Bahrain) |
| 85 | 1570825681 | An Optimized Ranking Based Technique towards Conversational Recommendation Models | Recommendations can be adjusted based on the liking of an individual by employing the technique of critiquing with conversational recommendation. For example, a product recommendation and a feature set are suggested to an individual. The individual can then either accept the suggestion or criticize it, producing a more refined suggestion. A modern embedding centered technique is incorporated into the recent model, latent linear critiquing (LLC). LLC aims to improve the embedding of the liking and critiques of an individual centered on particular product depictions (e.g., key phrases from individual feedbacks). This is achieved by exploring the arrangement of the embeddings to effectively improve the weightings following a linear programming (LP) design. In this paper, LLC is revisited. It has been observed that LLC is a grade centered technique which utilizes extreme weightings to enlarge estimated score gaps among favored and non-favored products. We observed that the final aim of LLC is the re-ranking rather than re-scoring. In this research article, an optimized ranking-based technique is proposed which aims to optimize embedding weights centered on noticed rank infringements from previous critiquing repetitions. The suggested model is evaluated on two recommendation datasets which comprise of individual feedbacks. Experimental outcomes reveal that ranking centered LLC usually performs better than scoring centered LLC and other standard approaches across diverse datasets, such as critiquing formats and several other performance measures. | http://dx.doi.org/10.12785/ijcds/150185 | Conversational Recommendation; Critiquing; Latent Linear Critiquing; Embedding | Sanjeev Dhawan (Department of Computer Science and Engineering, UIET, KUK, India); Kulvinder Singh (UIET, Kurukshetra University, India); Amit Batra (Kurukshetra University, India); Anthony Choi and Ethan Choi (Mercer University, USA) |
| 86 | 1570915727 | Evolutional Based Optimization Analysis for Three-element Control System | Simultaneous controller tunings for multiple loops are challenging because the best parameter settings for all loops have to be obtained in the concurrent analysis, whereby all parameters are affecting to each other's and the output variables. This paper purposely presents the accomplishment of Proportional-Integral-Derivative (PID) controller tunings by applying multi-objective optimization algorithm to a multiple loop known as three-element loop for both its servo and regulatory control objectives. Secondly, the research highlights the determination of lower limit (LL) and upper limit (UL) bounds by using necessity criterion of Routh-Hurwitz stability analysis. This research starts with empirical model identification and controller settings via classical techniques and multi-objective optimization approach. The Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) were selected to search the most compatible optimized controller tunings in the simulation of steam boiler drum function and the optimization analysis were conducted by a developed Graphical User Interface (GUI). On the whole, various simulation analysis was relatively compared by respective performance indexes and curve responses. Overall results showed that the optimization analysis proposed the controller settings corresponds better performances, and in particular, GA has performed better than PSO even through both methods are capable to suggest the highly satisfactory performances. | http://dx.doi.org/10.12785/ijcds/150186 | Multiple loop; Upper and lower bounds; Multi-objective optimization; Graphical user interface; Curve and indexes performances | Ing Ming Chew (Curtin University Malaysia, Malaysia); Filbert H. Juwono (Xi'an Jiaotong - Liverpool University, China); Wei Kitt Wong (Curtin University, Malaysia) |
| 87 | 1570917966 | Exploring Sentence Embedding Representation for Arabic Question/Answering | Question Answering Systems (QAS) are made to automatically respond with precise information to user questions that are phrased in natural language. Due to its intricate and rich morphology, Arabic QAS poses a significant problem. Information retrieval, text summarization, and question-answering systems all fall under the category of natural language processing activities where text representation is a critical step. Comparing SE representation to more traditional approaches like bag-of-words and word embedding, it has demonstrated encouraging results. In this study, we introduce a novel QA approach for the Arabic language that is based on passage retrieval and SE representation. It consists of three steps: "Question classification and query formulation", "Documents and passages retrieval", and then "Answers extraction". In this work, we adopt the AraBert pre-trained model to compute vector representation. It allows us to consider implicit semantics and the words' context within the text. Furthermore, in order to collect potential passages for user questions, we investigate a method for retrieving Arabic passages using the BM25 model, a query expansion process, and SE representation. The final answer is extracted by fine-tuning AraBERT parameters by ranking passages and extracting the most relevant ones. We carry out a number of tests with the CLEF and TREC datasets by following two different taxonomies. The outcomes demonstrate the efficacy of our methodology. | http://dx.doi.org/10.12785/ijcds/150187 | Question Answering; Arabic; NLP; Word Embedding; Sentence Embedding; AraBERT | Imane Lahbari and Said Ouatik EL Alaoui (Ibn Tofail University, Morocco) |
| 88 | 1570923150 | Implementing Image Processing and Deep Learning Techniques to Analyze Skin Cancer Images | Skin cancer is one of the few malignancies that seem to be completely incurable. If not diagnosed and maneuvered at the inauguration, it spreads to several components of the human body. It emerges while the cell of the skin is in contact with sunshine, and develops primarily due to the rattling progression of skin cells. To facilitate things simpler and more rapid rescue lives, a responsible automatic system for Identifying skin lesions is essential for earlier detection. The strategy for successful skin cancer detection employs image processing and Deep Learning technique. In this article, we have experimented with an amalgamation of various Image Processing Techniques including hair removal, median filter, Gaussian blur, and techniques of Deep Learning for the image of skin cancer study. The investigation of histopathology images desires a professional physician to categorize the images precisely. The Convolution Neural Network is the preliminary classifier utilized here. It is a cutting-edge learner operated in image classification since it can categorize images without depending on not automatic feature extraction from an individual image. The main objective of this investigation is to enhance the heftiness of the classifier by approximating a distinguishable combination of different Image Processing and Deep Learning Techniques. | http://dx.doi.org/10.12785/ijcds/150188 | Image Processing Technique; Convolutional Neural Networks(CNN; Hair Removal Technique | Snowber Mushtaq (National Institute of Technology, India); Omkar Singh (National Institute of Technology Srinagar, India) |
| 89 | 1570906087 | Comparison of YOLO (v3, v5) and MobileNet-SSD (v1, v2) for Person Identification Using Ear-Biometrics | The ear is a visible organ with a unique structure for each person. As a result, it can be used as a biometric to circumvent the constraints of person identification. Deep learning methods like You Only Look Once (YOLO) and MobileNet have recently significantly aided real-time biometric recognition. As a result, in this paper, we approach identifying a person using YOLOV3, YOLOV5, MobileNet-SSDV1, and MobileNet-SSDV2 deep learning algorithms using their ear biometrics. The used ear biometric is a standard dataset (EarVN1.0 Dataset) from 164 individuals with a total of 27,592 images. We chose 10 people at random, totaling 2057 pictures. Of these, 85% were used for training, 5% for validation, and 10% for testing. The performance of the algorithms is determined based on their accuracy and how smoothly the ear of a person is detected. The training accuracy of the algorithms is thresholded at 99.87%. MobileNet-SSDv1, MobileNet-SSDv2, YOLOv3, and YOLOv5 have testing accuracy that is 88%, 91%, 95%, and 96%, respectively. We concluded that the YOLOv5 model outperforms the others in terms of accuracy and size (16MB) for person identification using ear biometrics. | http://dx.doi.org/10.12785/ijcds/150189 | Person Identification; Ear Biometric; Deep Learning; YOLO; MobileNet; SSD | Shahadat Hossain (International University of Business Agriculture and Technology, Bangladesh); Humaira Anzum (Ahsanullah University of Science and Technology, Bangladesh); Md. Shamim Akhter Phd. (Ahsanullah University of Science and Technology (AUST), Bangladesh) |
| 90 | 1570974662 | RDMAA: Robust Defense Model against Adversarial Attacks in Deep Learning for Cancer Diagnosis | Attacks against deep learning (DL) models are considered a significant security threat. However, DL especially deep convolutional neural networks (CNN) has shown extraordinary success in a wide range of medical applications, recent stud-ies have recently proved that they are vulnerable to adversarial attacks. Adversari-al attacks are techniques that add small, crafted perturbations to the input images that are practically imperceptible from the original but misclassified by the net-work. To address these threats, in this paper, a novel defense technique against white-box adversarial attacks based on CNN fine-tuning using the weights of the pre-trained deep convolutional autoencoder (DCAE) called Robust Defense Model against Adversarial Attacks (RDMAA), for DL-based cancer diagnosis is introduced. Before feeding the classifier with adversarial examples, the RDMAA model is trained where the perpetuated input samples are reconstructed. Then, the weights of the previously trained RDMAA are used to fine-tune the CNN-based cancer diagnosis models. The fast gradient method (FGSM) and the project gra-dient descent (PGD) attacks are applied against three DL-cancer modalities (lung nodule X-ray, leukemia microscopic, and brain tumor magnetic resonance imag-ing (MRI)) for binary and multiclass labels. The experiment's results show that under attacks, the accuracy decreased to 35% and 40% for X-rays, 36% and 66% for microscopic, and 70% and 77% for MRI. In contrast, RDMAA exhibited substantial improvement, achieving a maximum absolute increase of 88% and 83% for X-rays, 89% and 87% for microscopic cases, and 93% for brain MRI. The RDMAA model is compared with another common technique (adversarial training) and outperforms it. Results show that DL-based cancer diagnoses are extremely vulnerable to adversarial attacks, even imperceptible perturbations are enough to fool the model. The proposed model RDMAA provides a solid foun-dation for developing more robust and accurate medical DL models. | http://dx.doi.org/10.12785/ijcds/150190 | Adversarial Attacks; Deep Learning; Cancer Diagnosis; Deep Convolutional Autoencoder | Atrab A. Abd El-Aziz (KafrelSheikh University, Egypt); Reda Abd Elwahab and Nour El Deen Mahmoud Khalifa (Cairo University, Egypt) |
| 91 | 1570982868 | A Parallel Approach of Cascade Modelling Using MPI4Py on Imbalanced Dataset | Machine learning is crucial in categorizing data into specific classes based on their features. However, challenges emerge especially in classification when dealing with imbalanced datasets in which the model exhibits bias towards the majority class. This research proposes a cascade and parallel architecture in the training process to enhance accuracy as well as speed compared to non-cascade and sequential respectively. This research will evaluate the performance of the SVM and Random Forest methods. The research finds that the Support Vector Machine (SVM) method with the Radial Basis Function (RBF) kernel notably increases accuracy by 1.25% over non-cascade classifiers. In addition, the use of Message Passing Interface for Python (MPI4Py) for training process across multiple cores or nodes proved that parallel processing significantly speeds up the training process up to 3.57 times faster than sequential training. These findings underscore the effectiveness of parallel processing in enhancing both the accuracy and efficiency of classification tasks in imbalanced data scenarios. | http://dx.doi.org/10.12785/ijcds/150191 | Cascade classifier; imbalanced data; MPI4Py; parallel processing; SVM | Suprapto Suprapto and Wahyono Wahyono (Universitas Gadjah Mada, Indonesia); Nur Rokhman (Gadjah Mada University, Indonesia); Faisal Dharma Adhinata (Institut Teknologi Telkom Purwokerto, Indonesia) |
| 92 | 1570988504 | Exploring the Impact of Shifting ERP Systems to the Cloud | The demand for improving efficiency, productivity, and business processes through cloud-based enterprise resource planning (ERP) has grown substantially in recent decades. This study aimed to investigate the benefits and challenges of moving traditional ERP to the cloud, as well as the success factors that facilitate the transformation process. To the best of our knowledge, few studies have explored these areas in depth. To achieve the specified aims, semi-structured interviews were conducted with government, semi-government, and private companies that had moved their traditional ERP to the cloud. The data were analyzed using thematic analysis. It showed that the movement to the cloud would bring significant cost savings, standardization of business processes, improved accessibility, mobility, and usability and facilitate rapid implementations and upgrades and strict security standards. However, organizations may face challenges during the transformation process, such as customization limitations, organization, and user resistance, concerns about Internet reliability and security issues. To ensure the success of the transformation process, the following typical success factors were identified: business process re-engineering, careful planning of the transformation approach, the quality of the project team, and effective communication. This paper provides valuable insights for businesses considering implementing or migrating to cloud-based ERP. | http://dx.doi.org/10.12785/ijcds/150192 | Cloud-based ERP; cloud ERP benefit; success factors; on-premise systems | Zahyah H Alharbi, ZA (King Saud University, Saudi Arabia); Norah J AlMouteq (SAMI Advanced Electronics, Saudi Arabia) |
| 93 | 1570991548 | Cyber Resilience Framework: Strengthening Defenses and Enhancing Continuity in Business Security | This study presents a comprehensive Cybersecurity Resilience Framework designed to fortify organizational defenses against the evolving landscape of cyber threats while enhancing business continuity. The aim is to provide businesses with a robust and adaptive strategy that extends beyond traditional cybersecurity paradigms. This study employs a methodology grounded in an extensive cybersecurity literature review to inform the conceptualization and iterative development of a resilient framework, integrating key elements from established sources and aligning with industry wisdom. By integrating governance and leadership principles, collaboration with external stakeholders, and continuous monitoring, the framework fosters a holistic approach to cyber resilience. Leveraging a behavioral perspective, the study explores human factors, user awareness, and decision-making processes, recognizing the critical role of organizational culture in fostering a cybersecurity-aware ethos. Findings reveal a roadmap that includes technology resilience, regular audits, and assessments, emphasizing evidence-based improvements. The framework addresses resource constraints, regulatory variability, and the dynamic threat landscape, promoting adaptability in the face of diverse organizational contexts. The significance of this study lies in its contribution to the ongoing evolution of cyber resilience strategies, offering organizations a practical guide to navigate the complexities of the digital realm. As businesses increasingly rely on interconnected technologies, this framework stands as a vital tool for enhancing security, safeguarding critical assets, and ensuring continuity in the face of an ever-changing cyber threat landscape. | http://dx.doi.org/10.12785/ijcds/150193 | Cyber Security; Threats; Risk Assessments; Resilience Framework; Business Security; Business Continuity | Ahmad Mtair AL-Hawamleh (Institute of Public Administration IPA, Saudi Arabia) |
| 94 | 1570991406 | Efficient 3D Instance Segmentation for Archaeological Sites Using 2D Object Detection and Tracking | This paper introduces an efficient method for 3D instance segmentation based on 2D object detection, applied to the photogrammetric survey images of archaeological sites. The method capitalizes on the relationship between the 3D model and the set of 2D images utilized to compute it. 2D detections on the images are projected and transformed into a 3D instance segmentation, thus identifying unique objects within the scene. The primary contribution of this work is the development of a semi-automatic image annotation method, augmented by an object tracking technique that leverages the temporal continuity of image sequences. Additionally, a novel ad-hoc evaluation process has been integrated into the conventional annotation-training-testing cycle to determine the necessity of additional annotations. This process tests the consistency of the 3D objects yielded by the 2D detection. The efficacy of the proposed method has been validated on the underwater site of Xlendi in Malta, resulting in complete and accurate 3D instance segmentation. Compared to traditional methods, the object tracking approach adopted has facilitated a 90% reduction in the need for manual annotations, The approach streamlines precise 3D detection, establishing a robust foundation for comprehensive 3D instance segmentation. This enhancement enriches the 3D survey, providing profound insights and facilitating seamless exploration of the Xlendi site from an archaeological perspective. | http://dx.doi.org/10.12785/ijcds/150194 | Underwater archaeology; AI; Convolutional Neural Network (CNN); 3D Instance Segmentation; Underwater photogammetry. | Maad Al-Anni (Al-Iraqia University, Iraq); Pierre Drap (Aix Marseille University, France) |
| 95 | 1570989432 | Meticulous review on Cutting-Edge Cervical Cancer cell Detection, Segmentation and Stratification of PapSmear Images using Image Processing and Machine learning Approach | Cervical cancer falls under the top most cancers found in women of developing countries since last many years. Classification of cervical cancer through a traditional microscopic approach is a monotonous and prolonged task. Most of the time hospital doctors cannot identify the cancer cells as sometimes the nucleus is difficult to see with naked eyes. Due to the different perspectives of doctors, cancer stages are classified falsely which leads to low recovery and late medication. The use of Image Processing and Machine Learning technologies can take off misclassification and inaccurate prediction. Although many deep learning techniques are available for cervical cancer cell detection and classification, performance of such techniques for prediction and classification with the real and sample dataset is the main challenge. In this paper, we did a thorough state-of-the-art review with the available current literature. The objective of this paper is to bring forth in-depth knowledge to novice researchers with the thorough understanding of the architecture of the computer assisted classification process. The current literature is studied, analyzed, and discussed with their approaches, results, and methodologies. | http://dx.doi.org/10.12785/ijcds/150195 | Image Processing; Machine Learning; Image Classification; Pattern Analysis; Feature evaluation; Feature selection | Barkha Bhavsar (KSV University, USA & LDRP Institute of Technology and Research, India); Bela Shrimali (Institute of Technology, Nirma University, India) |
| 96 | 1570914630 | A Novel Approach for Denoising ECG Signals Corrupted with White Gaussian Noise Using Wavelet Packet Transform and Soft-Thresholding | Purpose The electrocardiogram (ECG) is a popular technique for timely identification and assessment of heart problems. During recording, Various types of noise commonly contaminate ECG signals. which might lead to inaccurate diagnosis. When ambulatory patient monitoring and wireless recording are used, AWGN is one of the prominent noise that severely contaminates the ECG signal. Thus, precise ECG signals are essential for accurate heart disease diagnosis. Methods A wavelet denoising approach employing the Wavelet Packet Transform (WPT) is used to preprocess the obtained noisy data. WPT offers greater versatility due to more complete analysis of the signal. The Symlets 8 mother wavelet function is used to decompose ECG data into two levels. Using a two-level WPT decomposition, the noisy ECG signal is first divided into separate components including high- and- low frequency detail and approximation coefficients. The threshold value is then determined, and at every level, high frequency noise is effectively eliminated by applying thresholding technique to the detail coefficients, followed by the elimination of noise in the lower frequency range by thresholding the approximation coefficient. The inverse wavelet packet transform is used to rebuild the ECG signal from the retained coefficients Results Several tests were carried out to illustrate the effectiveness of the suggested strategy, and the results were compared with those of previously used signal denoising methods. When evaluated on the database of MIT-BIH, the suggested approach outperforms the existing state-of-the-art methodologies. Percentage root mean square difference (PRD), SNRimp, Signal to noise ratio (SNR) and Mean Squared Error (MSE) are used to assess the efficacy of denoising approach. Surrogated ECG signals with synthetic and gaussian white noise are employed. Conclusions Using the appropriate decomposition levels and mother wavelet, this proposed method developed a unique Wavelet Packet Transform approach for ECG signal denoising, in contrast to the already established techniques | http://dx.doi.org/10.12785/ijcds/150196 | Discrete Wavelet Transform (DWT); Wavelet Transform (WT); ECG; Noise reduction; denoising; Additive white Gaussian noise (AWGN) | Haroon Yousuf Mir (National Institute of Technology Srinagar Kashmir India, India); Omkar Singh (NIT Srinagar, India) |
| 97 | 1570987102 | The Interplay between Intellectual Capital, Business Intelligence Adoption, and the Decision to Innovate: Evidence from Jordan | This research explores the dynamic interplay among intellectual capital, the intention to adopt business intelligence (BI) technology, and the decision to innovate within the industrial landscape of Jordan. Using a quantitative approach, the study employs bootstrapping and PLS-SEM to analyze data from participants familiar with their companies' technological and innovation orientations. The findings reveal a noteworthy positive correlation between human capital and structural capital with the intention to adopt BI technology. Additionally, human capital demonstrates a significant positive association with the decision to innovate. The research further validates a positive relationship between the intention to adopt BI technology and the decision to innovate. The practical implications of these findings extend to decision-makers and managers in Jordan's industrial sector, underscoring the pivotal role of adopting business intelligence technology in fostering innovation. Significantly, by concentrating on innovation orientation in the Jordanian context, this paper contributes to the expanding body of research in developing countries. | http://dx.doi.org/10.12785/ijcds/150197 | Decision to Innovate; Business Intelligence; Intellectual Capital; Industrial Sector | Zaid Jaradat, ZJ (AlBayt University & School of Business, Jordan); Ahmad Mtair AL-Hawamleh (Institute of Public Administration IPA, Saudi Arabia); Marwan Altarawneh, Heba Hikal and Alya Elfedawy (Arab Open University, Saudi Arabia) |
| 98 | 1570989890 | AlphaFedForensic: Safeguarding Privacy and Enhancing Forensic Analysis in Federated Learning on Edge Devices | In this paper, a novel federated learning algorithm for decentralized settings on edge devices— AlphaFedAvg—is introduced. Using an adaptive learning rate approach based on Lipschitz and Smoothness parameters, AlphaFedAvg dynamically modifies the learning rate for every node. Through federated averaging, the approach accomplishes model aggregation, exhibiting enhanced convergence and performance. An extensive test configuration includes using Kali Linux to simulate network assaults, an ESP32 microcontroller connected to a laptop equipped with a sound sensor, and Wireshark and Scapy for traffic analysis. The Alpha algorithm offers a privacypreserving solution by effectively identifying and thwarting network attacks without gaining access to user data. The algorithm's performance is demonstrated in a comprehensive report generated. Evaluation against IID and non-IID datasets, such as Edge-IIoTset, and comparison with other models validate AlphaFedAvg's efficacy in federated learning applications. | http://dx.doi.org/10.12785/ijcds/150198 | Federated Learning; Privacy-Preserving; Anomaly Detection | Karam Salih (Ninevah University, Iraq); Najla Badie Aldabagh (Mosul University, Iraq) |
| 99 | 1570996865 | Approximate Computing-based Assistive Shopping Trolley for Visually Challenged People | Sighting is one of the five senses that help us to live in this world. According to the World Health Organization (WHO), more than 2 billion people face vision challenges that impose serious life problems. The visually challenged must utilize their senses, i.e., touch, smell, taste, and sound, to perform their daily tasks. Visually challenged people suffer from visual impairment, either partial or total. Their treatment could be corrective eyeglasses, assistive devices, and medical therapy. In this work, we target an assistive device, i.e., we propose, design and implement an approximate computing-based smart shopping cart for visually challenged people. The proposed assistive shopping cart utilizes the Internet of Things (IoT), e.g., Radio Frequency Identification (RFID) sensors, an Infrared (IR) Sensor, and an Arduino microcontroller. Audio was used during the shopping journey to tell the shopper his exact location on the mall, and the name and price of each item he scans. Combining many features in a single cart will result in a revolutionary device that might revolutionize the face of shopping for the visually impaired. The proposed cart improvements include obstacle detection, voice commands, and emergency assistance. The design cart emphasizes the potential of technology to bridge barriers and empower people with disabilities, as well as the beneficial influence that inclusive solutions may have on overall quality of life. | http://dx.doi.org/10.12785/ijcds/150199 | Approximate Computing; Assistive Technology; Internet of Things (IoT); Embedded System; Visually Impaired | Mahmoud Saleh Masadeh (Hijjawi Faculty for Engineering, Yarmouk University, Irbid, Jordan); Esraa Alkhdour, Heba AbuDiak and Rand Obaidat (Yarmouk University, Jordan) |
| 100 | 1570983167 | Microservices Adoption: An Industrial Inquiry into Factors Influencing Decisions and Implementation Strategies | Microservices Architecture (MSA) has emerged as a promising paradigm for building scalable and flexible software systems. While extensive research focuses on MSA's technical aspects, there remains a gap in understanding how practitioners make decisions to adopt and implement MSA in real-world organizational contexts. To address this gap, we conducted an in-depth qualitative study through 30 semi-structured interviews with experienced practitioners in the field. Our investigation unveils the intricate factors driving practitioners' decision-making processes during MSA adoption. We highlight the multifaceted influences of reusability, scalability, extensibility, maintainability, and other factors, shedding light on the motivations behind adopting MSA. Moreover, we delve into key strategies practitioners employ during MSA adoption, emphasizing the importance of the Modulith approach as a bridge between monolithic and MSA Our findings underscore the significance of practitioner experience in shaping MSA adoption decisions. | http://dx.doi.org/10.12785/ijcds/1501100 | Microservices Architecture; Modulith; Decision-Making; Adoption; Industrial Inquiry; Qualitative Study | Mehdi Ait Said, Sr. (Hassan First University, Morocco); Abdellah Ezzati (FST SETTAT, Morocco); Soukaina Mihi (University Hassan first of Settat, Morocco); Lahcen Belouaddane (Hassan First University, Morocco) |
| 101 | 1570948920 | Performance Evaluation of the "Incremental Conductance" and "Adaptive Hill Climbing" Maximum Power Point Trackers for Wind Energy Conversion System | A new design of Maximum Power Point Tracking (MPPT) algorithm is integrated towards the Wind Energy Conversion System (WECS) to grasp the ultimate possible power (Pmax). The controllers applied in this research are categorized in the form of what is called a Direct Power Control (DPC) used in checking the efficiency performance besides evaluating the controllers based on the uncertain rapid variations in the wind speed's profile. The aim of the research is to evaluate the efficiency and performance of the incremental conductance (INC) and the adaptive Hill-Climbing Search (HCS) for Wind Energy Conversion System (WECS) using uncertain rapid changes in the wind speed profiles. The Permanent Magnet Synchronous Generator (PMSG) has been chosen in the research modeling because of its reliability and robustness. The results derived from the execution of simulations revealed the fact that the speed of wind plays a crucial aspect in rotor's speed along the electromagnetic torque due to the relatively proportional relationship that takes place between the wind speed parameters and power factor. The controller is adapted to the WECS for analyzing the performance of the Incremental Conductance algorithm (INC), followed by adaptive Hill-Climbing Search (HCS). This study demonstrated that the adaptive HCS has shown high-efficiency performance under rapid changes in wind speed. As per the simulation of INC results, it shows a fast-tracking capability to stretch in approaching the peak maximum power point MPPT at the curve representing the power. | http://dx.doi.org/10.12785/ijcds/1501101 | Incremental Conductance algorithm; HCS; wind power; MPPT; PMSG; WECS | Ahmed Badawi (UDST, Qatar); Mario I Elzein (University of Doha for Science and Technology, Qatar); Hassan Ali (UDST, QATAR); Alhareth Mohammed Zyoud (Birzeit University, Palestine) |
| 102 | 1570995610 | Automatic Detection of COVID-19 from Chest X-Ray Images using EfficientNet-B7 CNN Model with Channel-wise Attention | Since the outbreak of the global COVID-19 pandemic in Wuhan, China, in 2019, its impact has been seen worldwide. Early detection of COVID-19 is very important, as it keeps the infected people isolated from other people, thus minimizing the risk of further transmission. The standard diagnostic approach is based on RT-PCR. However, due to the scarcity of PCR kits in some regions and the costs associated with this technique, there is a growing demand for alternative solutions. Recently, diagnosis of COVID-19 by medical imaging has been recognized as a valid clinical practice. Meanwhile, the massive increase in COVID-19 cases has put considerable pressure on radiologists responsible for interpreting these scans. This paper introduces an automated detection approach as a rapid alternative for COVID-19 diagnosis. We present a deep CNN model to differentiate between normal and pneumonia cases, as well as patients with COVID-19. Our approach is based on EfficientNet-B7 architecture and improved with Squeeze and Excitation block as an attention mechanism. In addition, we propose an innovative architecture that combines CNN with SVM to achieve the best performance. Experimental results show that the proposed framework provides better performance than existing SOTA methods, with an average accuracy of 97.50%, while the precision and recall of COVID-19 are both 100% without any pre- or post-processing. | http://dx.doi.org/10.12785/ijcds/1501102 | COVID-19; Chest X-ray; CNN; EfficientNet-B7; Squeeze and Excitation block; Support Vector Machine | Mohamed Rami Naidji (University of Djillali Liabes, Algeria); Zakaria Elberrichi (University Djillali Liabes, Algeria) |
| 103 | 1570990010 | Analyzing the Impact of Discretization Techniques on Real Time Simulation of DC Servomotor Using FPGA | This paper explains the strategies to create a hardware-based real time model of DC servomotor by utilizing the FPGA technology that can accurately simulate the behavior of the servomotor in real-time. Such FPGA based hardware model is useful for testing control algorithms, validating designs, and optimizing performance for various applications because of its reconfiguration capabilities. Continuous-time model is discretized using both Backward Euler (BE) and Trapezoidal (TRZ) methods for the real-time implementation on FPGA. The discretized models are coded using ‘C', converted to hardware descriptive language using Vivado high level synthesis tools, and the performance is analyzed with change in step-size by comparing with the transfer function (TF) model. With 100 µsec step-size, TRZ response is found to be matching with the TF model, however, a step-size of 0.6 µsec was required for the BE. Also analyzed the closed loop speed control performance of the hardware-based real time DC servomotor models with discrete PID controller, again by varying the step-size. Both the BE and TRZ models could track the reference speed within 2 msec, because of the PID controller, however faster dynamics was observed in case of TRZ as compared to BE, especially with larger step-size. These analysis shows the effect of step-size and the discretization technique for the real-time modeling, however, with a suitable selected values, the developed FPGA model can be utilized efficiently for the development of suitable control algorithm. | http://dx.doi.org/10.12785/ijcds/1501103 | Backward Euler; DC Servomotor; FPGA; High Level Synthesis; Discrete PID Controller; Trapezoidal | Mini K Namboothiripad (Agnel Charities FCRIT Vashi, India) |
| 104 | 1570981963 | Process and Impact Evaluation of Artificial Intelligence in Managerial Accounting: A Systematic Literature Review | This systematic literature review aims to examine how researchers have evaluated the role of Artificial Intelligence (AI) in Management Accounting (MA), focusing on process and impact evaluations. The findings of the nine recent research papers completed between 2019 and 2023 were incorporated in this review. The study followed a three-phase, structured process: planning the review, conducting it, and reporting the findings. The study utilized multiple databases to source relevant papers and applied stringent inclusion and exclusion criteria, followed by a thematic synthesis approach. The results showed that the reviewed articles highlighted four aspects of AI in MA process evaluation: technological acceptance and usability, ethical and security considerations, skills and competence, and the decision-making process. Two aspects were pinpointed under impact evaluation in the reviewed articles: enhancement of accounting practices and the evolution of roles and skills. Two important predictions that the papers repeatedly asserted were that AI adoption in MA is still in its early stages but will transform the field soon. It will dramatically enhance MA practices, simultaneously creating new challenges and skills requiring urgent attention. Ethical and security considerations were also stressed in the reviewed articles. As AI algorithms are developing rapidly, the findings are limited to only giving a broad picture of the recent state of the field. However, this paper identified aspects of AI in MA evaluation that should contribute to setting evaluation criteria, which would be useful for future assessments. In light of the high potential of AI in the field of MA, this paper contributes to developing a comprehensive overview of the use of AI in this field. Furthermore, systematic reviews on AI applications in the MA field are limited. Therefore, this research addresses this gap by systematically reviewing recently published research papers. | http://dx.doi.org/10.12785/ijcds/1501104 | Managerial Accounting; Artificial intelligence Data Privacy; Business Intelligence; Decision-making; Machine learning; Ethical considerations | Dawla Almulla (University of Bahrain, Bahrain); Mohamed Abbas (Electricity & Water Authority, Bahrain); Adel Ismail Al-Alawi (University of Bahrain, Bahrain); Lamya Alkooheji (Bahrain) |
| 105 | 1570871127 | Semi-symmetrical Coprime Linear Array with Reduced Mutual Coupling Effect and High Degrees of Freedom | Direction finding based on the coprime array has received a lot of interest because of the increased degrees of freedom and larger aperture size compared with the uniform linear array. In this paper, a new semi-symmetric coprime array (SSCA) configuration is proposed to reduce the mutual coupling impact for passive sensing by exploiting the relocation of the redundant elements. First, we examine the difference co-array of the prototype coprime array (PCA) and identify the redundant elements. The second subarray (N-subarray) is then split into two subarrays on opposite axis, and the zero-lag location is moved to a new location to construct large contiguous lags in the difference co-array. The investigation of the SSCA is performed. The analytical expressions of the contiguous lags, unique lags, and aperture size are derived. The numerical and simulation results show that the proposed SSCA array can achieve less mutual coupling leakage as compared to other array types | http://dx.doi.org/10.12785/ijcds/1501105 | DOA; Coprime array; Mutual coupling; Difference co-array | Fatima A. Salman (Al-Nahrain University, Iraq); Bayan Mahdi Sabbar (University of Al-Mustaqbal, Iraq) |
| 106 | 1570955906 | Rocking Across Borders: An Analysis of the Musical Differences between Bangladesh and West-Bengal Rock Songs using Spotify Audio Features | Over the last few decades, there has been a significant increase in the availability and utilization of large music collections. However, most studies of these collections have been limited to Western music, which hinders our ability to comprehend the diversity and commonality of music across all cultures. Based on popularity from Spotify, it has been discovered that Bangladeshi rock music is more popular than the rock music from West Bengal, India. Previous research suggests that listeners from diverse cultural backgrounds may have varying preferences when comes to music appreciation. This research aimed to explore the reasons behind the popularity of Bangladeshi rock music compared to West Bengal rock music. By extracting various features from songs, the study sought to identify what is the reason behind the popularity of Bangladeshi rock music and whether there are any differences in terms of musical features. | http://dx.doi.org/10.12785/ijcds/1501106 | Data Mining; Musicological Analysis; Cultural Diversity; Feature Extraction; Hypothesis Testing; Pearson Correlation | Mohammed Raihan Ullah, Moshiur Rahman Autul, Durjoy Dey and Partha Protim Paul (Shahjalal University of Science and Technology, Bangladesh) |
| 107 | 1570980501 | A Survey on the Machine Translation Methods for Indian Languages: Challenges, Availability, and Production of Parallel Corpora, Government Policies and Research Directions | Since 1991, machine translation has been a prominent research area in India, with IIT Kanpur pioneering the original work which has since been expanded to several universities. Only 10 percent of India's 1.3 billion inhabitants can read, write, and speak English with varying degrees of competence, which makes machine translation crucial in overcoming the linguistic barrier to the internet. The Indian market for commercial products and events is greatly influenced by local languages, making the development and translation of region-based content an essential research topic nowadays. However, Indic-to-Indic language direct translation has faced several challenges and is still going through the experimental phase. Several government-sponsored projects are being undertaken in this regard. Still, there are limited sentence-aligned parallel bi-text resources available for the majority of Indian language pairs. This paper presents a detailed survey of the current trends of research on machine translation between Indian languages, along with their challenges over time. It also presents a timeline of recent research conducted and key findings of past surveys conducted over a decade. Under a single canopy, this paper provides sources of data, the progress made in developing datasets for low-resource Indian languages, various models of translation, encouragement from Indian Govt., and finally, new research directions. | http://dx.doi.org/10.12785/ijcds/1501107 | Machine Translation; RBMT; SMT; NMT; Low-Resource Indian languages; BLEU | Sudeshna Sani (Woxsen University, India & Koneru Lakshmaiah Education Foundation, India); Samudra Vijaya (Koneru Lakshmaiah Education Foundation, India); Suryakanth Gangashetty (Koneru Lakshmaiah Education Foundation Vaddeswaram AP, India) |
| 108 | 1570989122 | Attendance System Optimization through Deep Learning Face Recognition | The significance of face recognition technology spans across diverse domains due to its practical applications. This study introduces an innovative face recognition system that seamlessly integrates Multi-task Cascaded Convolutional Neural Networks (MTCNN) for precise face detection, VGGFace for feature extraction, and Support Vector Machine (SVM) for efficient classification. The system demonstrates exceptional real-time performance in tracking multiple faces within a single frame, particularly excelling in attendance monitoring. Notably, the "VGGFace" model emerges as a standout performer, showcasing remarkable accuracy and achieving an impressive F-score of 95% when coupled with SVM. This underscores the model's effectiveness in recognizing facial identities, attributing its success to robust training on extensive datasets. The research underscores the potency of the VGGFace model, especially in collaboration with various classifiers, with SVM yielding notably high accuracy rates. | http://dx.doi.org/10.12785/ijcds/1501108 | Attendance; Face Detection; Face Recognition; Feature Extraction; Transfer Learning | Mahmoud Emad Aldin Ali (Marwadi University & Syrian Virtual University, India); Anjali Diwan (Marwadi University, India); Dinesh Kumar (Bennett University, India) |
| 109 | 1570989271 | Diagnosing Autism Spectrum Disorder using a Smart Model | The brain development and physical appearance of the face are both impacted by the neurologic disorder known as Autism Spectrum Disorder (ASD). Children with ASD exhibit different facial landmarks from children who normally grow with Typical Developing (TD), despite the fact that the disorder is thought to be inherited. When a child's behavioral traits and facial features are examined, the likelihood of an accurate diagnosis is highest. In this work, we provide a deep learning model-based method for diagnosing autism that divides children into two categories: possibly healthy and potentially unhealthy. The Tesorflow and Keras libraries are used by the suggested deep learning model to carry out feature extraction and picture classification. A dataset obtained from the Kaggle repository is used to train and evaluate the model. the dataset that was used to test this model consisted of 2,122 that is excluded from the original 2,940 images due to the quality and race. The testing of the proposed model results in an Area Under Curve (AUC) with 99.8% and accuracy of 97.3%. This model proves its high diagnosis accuracy, ease of use, and fast decision. | http://dx.doi.org/10.12785/ijcds/1501109 | Autism Spectrum Disorder; Convolutional neural network; Deep learning; Keras; Tensorflow | Mazin Swadi (Deep Learning, Iraq); Muayad Sadik Croock (University of Technology, Iraq) |
| 110 | 1570991633 | Revolutionizing Cloud-Based Task Scheduling: A Novel Hybrid Algorithm for Optimal Resource Allocation and Efficiency in Contemporary Networked Systems | Need for cloud computing has increased in the age of contemporary networked systems, driving the pursuit of optimal resource allocation and data processing. This is especially important in essential fields where security depends on computing performance, such as transportation systems. Even after much research has been done on the management of resources in cloud computing, finding algorithms that maximize job completion, minimize costs, and maximize resource consumption has remained a top priority. However, existing techniques have shown limitations, which calls for new ways. This is our work, which shown the novel hybrid approach that has the potential to completely change the game. The Neural Network Task Classification (N2TC) is the result of the merging of neural networks with genetic algorithms. This ground-breaking method skillfully applies the Genetic Algorithm Task Assignment (GATA) for resource allocation while utilizing neural networks for task categorization. Notably, our algorithm carefully considers execution time, response time, costs, and system efficiency in order to promote fairness, a defense against resource scarcity. Our method achieves a remarkable 13.3% cost reduction, a stunning 12.1% increase in response time, and a 3.2% increase in execution time. These strong indicators act as a wake-up call, announcing the power and revolutionary potential of our hybrid algorithm in transforming the paradigms around cloud-based task scheduling. This work represents a turning point in cloud computing, demonstrating an innovative combination of algorithms that not only overcomes current constraints but also ushers in a new era of efficacy and efficiency that has far-reaching implications outside the domain of transportation systems. | http://dx.doi.org/10.12785/ijcds/1501110 | Cloud computing; Task Scheduling; Resource Allocation; Neural Network; Genetic Algorithm | Punit Mittal and Satender Kumar (Quantum University, India); Swati Sharma (Meerut Institute of Engineering and Technology, India) |
| 111 | 1570997799 | Exploring Radio Frequency-Based UAV Localization Techniques: A Comprehensive Review | UAVs (unmanned aerial vehicles) and WSNs (wireless sensor networks) are now two well-established technologies for monitoring, target tracking, event detection, and remote sensing. Typically, WSN is made up of thousands or even millions of tiny, battery-operated devices that measure, gather, and send information from their surroundings to a base station or sink. Within the realm of wireless positioning and communication, UAVs have garnered a lot of interest because of their remarkable mobility and simplistic deployment to tackle the problems of imprecise sensor placement, inadequate infrastructure coverage, and the massive quantity of sensing data that WSN collect. A crucial prerequisite for many position-based WSN applications is node location or localization. The usage of UAVs for localization is more preferable than permanent terrestrial anchor nodes due to their high accuracy and minimal implementation complexity. The possible interference or signal block in such operating environment, however, might cause the Global Positioning System (GPS) to become ineffective or unobtainable. In these conditions, the need for innovative UAV-based sensor nodes location technologies has become essential. Radio frequency (RF) based localization techniques is reviewed in the current paper. We examine the available RF features for localization and look into the current approaches that work well for unmanned vehicles. The most recent research on RF-based UAV localization is reviewed, along with potential avenues for future investigation. | http://dx.doi.org/10.12785/ijcds/1501111 | Unmanned Aerial Vehicles; Wireless Sensor Networks; Localization; Radio Frequency | Suha Abdulzahra (University of Babylon, Iraq); Ali Kadhum M. Al-Qurabat (University of Babylon & College of Science for Women, Iraq) |
| 112 | 1570913433 | Towards Robust and Low Latency Security Framework for IEEE 802.11 Wireless Networks | The vulnerability in wireless networks require additional security, integrity, and authentication. The outdated "Counter Mode Cipher Block Chaining Message Authentication Code Protocol" (CCMP) has lately taken the place of the flawed "Wired Equivalent Privacy" (WEP) protocol for authenticating IEEE 802.11 wireless local area networks (WLANs). IEEE 802.11s, a draught standard for wireless mesh networks, also recommended using CCMP (WMNs). Due to CCMP's two-pass operation, multi-hop wireless networks like WMN have a considerable latency problem. An increase in latency results in a decrease in service quality for real-time multimedia applications sensitive to delays. In addition to highlighting the CCMP's vulnerability to pre-computation time-memory trade-off (TMTO) attacks, this paper recommends improving WLAN packet security by implementing a per-packet security mechanism. Furthermore, suggested is a fresh, dependable, low-latency foundation for WMN. The security framework's architecture employs a piggyback challenge-response mechanism to ensure data secrecy and data integrity ۔The use of a secret nonce, a new encryption key for each packet, and packet-level authentication are all features of the Piggyback challenge-response protocol. By authenticating every packet, unauthorized access can be swiftly prevented. | http://dx.doi.org/10.12785/ijcds/1501112 | Information Security; Social Network; Decentralization; Individual Servers; Decentralized Framework | Inam Ul Haq (2 KM Renala Bypass, Renala Khurd, Okara & University of Okara, Pakistan); Amna Nadeem (Lahore Garrison University, Pakistan); Yasir Ahmad (University of Okara, Pakistan); Saba Ramzan (Sharif College of Engineering Lahore, Pakistan); Nazir Ahmad (University of Okara, Pakistan) |
| 113 | 1570978987 | Blockchain Technology and Virtual Asset Accounting in the Metaverse: A Comprehensive Review of Future Directions | This research focuses on the metaverse's evolving trend and the potential application of blockchain technology in the accounting of virtual assets in this digital domain. The metaverse introduces a new economy in which users may earn real-world revenue through virtual activities, necessitating the need for efficient and dependable virtual asset accounting. Blockchain technology, with its decentralized and immutable record, appears to be a viable answer to these problems. This paper discusses the present status of blockchain technology for accounting for virtual assets in the metaverse as well as its potential role for businesses and the economy. It also determines the technology's issues and limits and makes recommendations for further development. The approach of this study is based on a comprehensive review of the existing literature on the interactions between blockchain technology, virtual asset accounting, and the metaverse. The findings indicate that blockchain technology has the potential to transform virtual asset accounting in the metaverse by improving security, transparency, and consistency. However, scalability and legal/regulatory issues must be overcome before it can completely achieve its promise. Accounting experts, developers, and stakeholders interested in the convergence of blockchain technology and the metaverse economy will find this paper useful. | http://dx.doi.org/10.12785/ijcds/1501113 | Metaverse; Blockchain Innovation; Blockchain-based Accounting; Virtual Asset Accounting | Ahmad Mtair AL-Hawamleh (Institute of Public Administration IPA, Saudi Arabia); Marwan Altarawneh, Heba Hikal and Alya Elfedawy (Arab Open University, Saudi Arabia) |
| 114 | 1570922929 | Analytical comparison on detection of Sarcasm using machine learning and deep learning techniques | Sentiment Analysis is used in Natural Language processing to detect the opinion of the text/sentence put in by the user. A lot of challenges are faced while detecting the sentiment and one of them is the presence of sarcasm. Sarcasm is very difficult to detect and there could be ambiguity about the presence or absence of sarcasm. Various rule based methods have been used in the past by researchers to detect sarcasm. However, the results have not been promising. The models developed using machine learning classifiers have gained popularity over the statistical and rule based methods. Recently, deep learning techniques have been popularly used to detect the presence of sarcasm. In this paper, we have used eight machine language classifiers such as Naïve Bayes, Support Vector Machine, etc. to detect sarcasm. Deep learning techniques are also been used along with the machine learning techniques. An ensemble model has also been trained and tested on both the datasets. Bidirectional Encoder Representations from Transformers technique has given the best performance among the deep learning and machine learning techniques with an accuracy score of 92.73% and f-score of 93% on the news headlines dataset and an accuracy score of 75% and f-score of 74% on the reddit dataset. | http://dx.doi.org/10.12785/ijcds/1501114 | Sarcasm Detection; Machine Learning; Ensemble model; Deep Learning; Social media | Ameya Parkar and Rajni Bhalla (Lovely Professional University, India) |
| 115 | 1570958321 | Phishing Website Classification using Machine Learning with Different Datasets | The classification of phishing websites through the analysis of their URLs is a technique used to enhance the capabilities of systems designed to detect malicious websites. However, the evolution of phishing sites has allowed them to achieve higher levels of sophistication, making proactive detection more complex. The central focus of this article revolves around the exploitation of deep learning models and machine learning techniques with lexical analysis of their URLs to facilitate the classification, detection, and preventive mitigation of phishing websites. Our study includes the evaluation of a selection of commonly castoff machine learning algorithms, specifically Random Forest, K-Nearest Neighbors, Support Vector Machines, Gradient Boosting, Decision Tree, Bagging, AdaBoost and ExtraTree, as well as the deep neural network model. To assess the effectiveness of these algorithms and models, we conduct our analysis using two distinct URL datasets, one from 2016 and the other from 2021. Through lexical analysis, we extract significant features from the URLs and then calculate the accuracy of each algorithm on both datasets. Our results reveal that some algorithms achieve remarkable accuracy scores of up to 99% when applied to the 2016 dataset. However, this score decreases to less than 91% when applied to the dataset collected in 2021. | http://dx.doi.org/10.12785/ijcds/1501115 | Phishing; URL; Classification; Machine Learning; Deep Learnig | Habiba Bouijij (University of Mohammed V & ENSIAS, Morocco); Amine Berqia (University of Mohammed V, Morocco) |
| 116 | 1570958610 | Real-time Speech-based Intoxication Detection System: Vowel Biomarker Analysis with Artificial Neural Networks | Alcohol consumption can have impacts on the voice, and excessive consumption can lead to long-term damage to the vocal cords. A new procedure to automatically detect alcohol drinkers using vowel vocalizations is an earlier and lower-cost method than other alcohol drinker-detecting models and equipment. The hidden parameters of vowel sounds (such as frequency, jitter, shimmer, harmonic ratio, etc.) are significant for recognizing individuals who drink or do not drink. In this research, we analyze 509 multiple vocalizations of the vowels (/a, /e, /i, /o, and /u) from 290 multiple records of 46 drinkers and 219 multiple records of 38 non-drinkers. The age group is 22 to 34 years. Apply the 10-fold cross-validation vowelized dataset on intelligent machine learning models and incremental hidden layer neurons of artificial neural networks (IHLN-ANNs) with Backpropagation. The findings showed that experimental ML models such as Naïve Bayes (NB), Random Forest (RF), k-NN, SVM, and C4.5 (Tree) performed well. The RF model performed best, with 95.3% accuracy. We also applied the incremental hidden layer (HL) neurons BP-ANNs model (from 2 to 5). In this analysis, accuracy increased proportionally with the incremental neurons (2-5) in the HL of the ANN. Now of 5 neurons HL ANN, the model performed with a highly accurate 99.4% without an over-fit problem. It will implement smartphone apps for caution and alerts for alcohol consumers to avoid accidents. Voice analysis has been explored as a non-invasive and cost-effective means of identifying alcohol consumers. | http://dx.doi.org/10.12785/ijcds/1501116 | Alcohol Consumers; Voice Parameters; Machine Learning; Neural Networks; ANN | Terlapu Panduranga Vital (jntuK, India & Aditya Institute of Engineering Technology and Management, India); Ram Prasad Reddy Sadi (Anil Neerukonda Institute of Technology and Sciences, India) |
| 117 | 1570965876 | Performance Evaluation of Deep Learning Models for Face Expression Recognition | Face expression recognition is a crucial computer vision task with versatile applications. In this study. Explore the challenges and solutions Real-time emotion recognition is an important research problem. To solve this problem, transfer learning was used. we assess the performance of five deep learning models: CNN, VGG16, Inception V3, MobileNet V2, and DenseNet121. We evaluate their accuracy during training, validation, and testing, revealing their strengths and weaknesses. CNN delivered consistently high performance with 98% accuracy in testing, showcasing its efficacy in facial expression recognition. VGG16, a popular pre-trained model, achieved 97% accuracy in testing, making it a robust choice. Inception V3 demonstrated exceptional accuracy, with 99% in testing, highlighting its ability to capture intricate facial expression patterns. MobileNet V2, known for its efficiency, achieved 100% accuracy in testing, making it precise and computationally efficient. DenseNet121 yielded commendable results with 96% accuracy in testing, offering reliable performance. MobileNet V2 and Inception V3 particularly excelled with perfect or near-perfect accuracy across all phases. In this article CK+ Dataset used that consist of 981 images distributed into seven emtions. These findings aid researchers and developers in selecting suitable models based on their specific requirements, considering both accuracy and computational efficiency. Further fine-tuning may enhance these models for real-world applications. | http://dx.doi.org/10.12785/ijcds/1501117 | Face Expression; Emotion Recognition; Deep Learning; Transfer learning | Raed Ibrahim Khalil, Sr (University of Almustaniriya, Iraq); Abbas Hussien Miry (Mustansiriyah University, Iraq); Tariq Mohammed Salman (Mustansiriyah University & College of Engineering, Iraq) |
| 118 | 1570997351 | DecentraLend: A Blockchain-based Monetization for Decentralized Lending System | The concept of borrowing or lending or renting goods or tools from others is commonly based on the centralized distribution, which means that the transaction of items takes place between a lender and a borrower. The popularity of the handheld digital devices among mass people and the availability of the secured tracking technologies, such as blockchain, bring the opportunity to introduce a new concept of the decentralized virtual lending or rental system. Since the blockchain eases the financial transactions secured and traceable for individuals, a lent item can be tracked by relating the possessor of that item to a financial transaction. In this paper, we proposed a decentralized virtual lending system based on blockchain for lending and borrowing physical items among individuals or companies. Our proposed system incorporates the monetary value of a physical item in the blockchain and tracks the current possessor and ensure safety of the lent item. Moreover, our proposed decentralized virtual lending system incorporated a recommendation mechanism for the users to borrow an item from the list of best alternatives without visiting a traditional rental company or the owner and allows an individual or a company to monetize by lending their goods and tools to others. | http://dx.doi.org/10.12785/ijcds/1501118 | Decentralized app; Blockchain app; Decentralized monetization; Decentralized recommendation | Mostofa Kamal Rasel, Mahamudul Hasan, Mohammad Rifat Ahmmad Rashid, Md Hasanul Ferdaus and Mohammad Manzurul Islam (East West University, Bangladesh); Md Ali (Universiti Technology Malaysia, Malaysia); Maheen Islam (University of Dhaka & East West University, Bangladesh); Md Ariful Islam (East West University, Bangladesh) |
| 119 | 1570998404 | Human Detection in Clear and Hazy Weather Based on Transfer Learning With Improved INRIA Dataset Annotation | Human detection plays a pivotal role in many vision-based applications. Effectively detecting humans across diverse environments and situations significantly contributes to enhancing human safety. However, this effectiveness encounters challenges, particularly in hazy conditions that reduce visibility and blur images, thereby impacting the accuracy of existing detection algorithms. Additionally, the quality of dataset annotations significantly affects the accuracy of these systems. Poor annotations lead to insufficient training of detection models, resulting in higher error rates and reduced efficacy in real-world scenarios. To tackle these challenges, we've introduced new, more precise annotations for the INRIA dataset. These enhancements overcome limitations within the dataset, particularly instances where numerous individuals in images lacked proper labeling. This augmentation aims to improve training robust detection models and provide a more accurate evaluation of the model's performance. Our experiments have yielded notable improvements, showcasing a 20.37% increase in Average Precision and a substantial 68.19% reduction in False Negatives. Moreover, we've developed a deep-learning model for human detection, leveraging transfer learning to fine-tune the YOLOv4 model. Experimental results demonstrate that our proposed model accurately detects pedestrians under various weather conditions, including both clear and hazy scenarios. It achieves high average precision and F-Scores while maintaining efficient real-time operation at 55.4 FPS. These advancements significantly enhance the reliability and applicability of human detection systems. | http://dx.doi.org/10.12785/ijcds/1501119 | Human Detection; CNN; Deep Learning; YOLOv4; Transfer Learning; INRIA Dataset | Yassine Bouafia (University of Batna2 & LaSTIC Laboratory, Algeria); Larbi Guezouli (HNS-RE2SD, Batna, Algeria & LEREESI Laboratory, Algeria); Hicham Lakhlef (Heudiasyc, University of Technology of Compiègne, France) |
| 120 | 1570956161 | Mitigating Data Variability and Overfitting in Deep Learning Models for Atrial Fibrillation Detection Using Single-Lead ECGs | Despite the growing potential of deep learning in diagnosing Atrial Fibrillation (Afib), challenges such as overfitting and limited generalizability continue to persist. These limitations are accentuated in single-lead ECGs generated from wearable devices, which frequently suffer from inadequate annotation and substantial data variability. This study seeks to address these challenges by enhancing both the accuracy and generalizability of Afib detection algorithms. We introduce Afib-CNN, a specialized Convolutional Neural Network engineered for 9-second, single-lead ECGs. The architecture comprises ten convolutional blocks and three fully connected layers, focusing on computational efficiency. To mitigate data variability, we apply advanced pre-processing techniques like Moving Average by Convolution Filter (MAConv) and Minimum-Maximum Normalization. Further dataset refinement is achieved using z-score normalization and a shifted-length overlapping technique. The effectiveness of our model is rigorously validated across three distinct ECG databases, demonstrating robust intra- and inter-patient generalizability. Employing 10-fold stratified cross-validation, Afib-CNN exhibits exemplary performance, achieving mean F1 scores of 98\%, 97\%, and 99\% on the CinC2017, CPSC2018, and MIT-AFIB datasets, respectively. The model also attains an F1 score of 98\% on the CinC2017 test set. Comparative analyses demonstrate that Afib-CNN successfully balances high performance, computational efficiency, and robust generalization. These characteristics render it well-suited for practical clinical deployment. | https://dx.doi.org/10.12785/ijcds/1501120 | Convolutional neural network (CNN) , Arrhythmia classification , Short single-lead ECG recordings , ECG Data Variability , Overfitting , Wearable ECG. | Khadidja Benchaira (University Mohamed Khider & LESIA Laboratory, Algeria); Salim Bitam (University of Biskra & LESIA Laboratory, Algeria); Zineb Djihane Agli (University of Mohamed Khider Biskra, Algeria) |
| 121 | 1570954859 | Hybrid K-means and Principle Components Analysis (PCA) for Diabetes Prediction | Diabetes is the "silent killer," stealing the lives of millions of people worldwide. There are many reasons for diabetes, such as increasing glucose, Cholesterol, Systolic BP, and age. These are considered the four main reasons that cause diabetes. The challenge in diabetes is how to predict the human illness early to start treatment immediately after discovering diabetes; This can be the most challenging thing in diabetes discovery because tens of features may cause diabetes. This thesis proposes a model consisting of data mining and Machine Learning (ML) algorithms to predict if humans can have diabetes or not in the future. the prediction is made up of compensating two datasets one dataset is used to reconfirm the other dataset in making more accurate prediction; This can be performed using the k-means-PCA hybrid model and the highest weight selection of features that widely cause diabetes. The selected features help the ML algorithm predict the model's accuracy, which indicates the prediction model's accuracy. Simulation results show that the predict-diabetic patients increased from 53 from the original datasets to 142 after applying the proposed model. Simulation outcomes also prove that the Random Forest ML model gives the highest accuracy of other ML models, reaching 95.2%. | http://dx.doi.org/10.12785/ijcds/1501121 | Artificial intelegence; Machine Learning; Healthcare; Diabetes | Ahmed Abed Mohammed (Islamic University, Najaf, Iraq); Putra Sumari (Universiti Sains Malaysia (USM) & School of Computer Science, Malaysia); Kassem Attabi (The Islamic University, Iraq) |
| 122 | 1570971343 | Reptile Search Algorithm for Association Rule Mining | Association rule mining (ARM) is a very popular, interesting, and active research area in data mining. It aims to find potentially useful connections between different attributes in a defined dataset. ARM is an NP-complete problem and become a fertile field for optimization applications. The Reptile Search Algorithm (RSA) stands out as an innovative bio-inspired meta-heuristic approach. It draws inspiration from the hunting and encircling behaviors exhibited by crocodiles in nature. It is a well-known optimization technique for solving NP-complete issues. Since its introduction by Abualigah et al. in 2022, the approach has attracted considerable attention from researchers and has extensively used to address diverse optimization issues in several disciplines. This is due to its satisfactory execution speed, efficient convergence rate, and superior effectiveness compared to other widely recognized optimization methods. This paper presents new version of the reptile search algorithm for resolving the association rules mining challenge. Our proposal inherits the trade-off between local and global search in optimization issues that the Reptile search algorithm demonstrated. In order to demonstrate the power of our proposal, a series of experiments are cried-out on a varied well-known, employing multiple comparison criteria. The outcomes show a clear superiority of the proposed approach in-face-of famous association rules mining algorithms in terms of CPU time, fitness criteria, and the quality of generated rules. | http://dx.doi.org/10.12785/ijcds/1501122 | Data mining; Association rule mining; Bio-Inspired Approaches; Reptile Search Algorithm | Abderrahim Boukhalat (University of M'Sila & LIM Laboratory, University of Souk Ahras, Algeria); Heraguemi KamelEddine (National School of Artificial Intelligence, Algeria); Mohamed Benouis (M'sila University, Algeria); Brahim Bouderah (Pro, Algeria); Samir Akrouf (University Mohamed Boudiaf MSILA, Algeria) |
| 123 | 1570972516 | Mapping crop types at a 10 m scale using Sentinel-2 data and machine learning methods | Crop classification plays a vital role in crop status monitoring, crop area estimation and food production. Remote sensing data is widely accepted for crop classification at the remote location. However, crop classification is challenging due to spectral and spatial similarities, complex land structure, temporal inconsistencies, and environmental parameters. Therefore, in the present study, an effort has been made to identify, classify and map multiple crops from the complex environment using the Sentinel-2 dataset and advanced machine learning methods such as random forest, Spectral Angle Mapper (SAM), Maximum Likelihood Classifier (MLC), K-means clustering and Iterative Self-Organizing Data Analysis Technique (ISODATA). The crop spectral features were identified using the Normalized Difference Vegetation Index (NDVI). The NDVI outcomes ranged between -0.91 and 0.54, which were then used to identify crop areas. Ground reference data, Google map, and Google Earth data were used to determine the crop classes, train the data, and validate the results. The five major crops viz. Cotton, Paddy, Orchard, Yellow split Pigeon peas, Chickpeas and Other crops were identified and classified efficiently. According to the experimental results, the random forest approach had the best overall accuracy, 87.71% and a kappa value of 0.86 than other methods. Alternatively, the ISODATA method provided overall accuracy of 85.01% with a kappa value of 0.82. The agricultural decision-makers can use the results of this study for decision-making and management. | http://dx.doi.org/10.12785/ijcds/1501123 | Crop types mapping; Sentinel-2 data; Machine learning; Random Forest; NDVI | Atiya Khan (G H Rasisoni College Engineering Nagpur, India); Chandrashekhar Himmatrao Patil (Vishwanath Karad MIT World Peace University, India); Amol D. Vibhute (Symbiosis International Deemed University Pune, India); Shankar Mali (Vishwanath Karad MIT World Peace University, India) |
| 124 | 1570977603 | Efficient Heuristics for Timetabling Scheduling in Blended Learning | Blended learning, an innovative educational approach adopted by the Saudi Electronic University, combines traditional face-to-face teaching with online methods to improve educational outcomes. However, implementing it effectively faces challenges, especially in shifting physical classes to virtual formats for faculty and students spread across distant campuses, potentially compromising the core principles of blended learning and reducing its effectiveness. Ideally, each student would attend one face-to-face and one virtual session per course weekly. Yet, operational challenges emerge due to geographical disparities between faculty and student locations, necessitating the conversion of face-to-face sessions into virtual ones. To address these challenges, this study proposes two novel heuristic strategies for scheduling timetables in blended learning contexts. The first heuristic, called Minimum Load Accumulation Heuristic (MLAH), aims to evenly distribute teaching loads among faculty members and time slots while maximizing the number of groups assigned to faculty members from the same campus. The second heuristic, called Average Load Accumulation Heuristic (ALAH), calculates the average load of all faculty members and time slots and reduces the number of iterations searching for the minimum load, as performed by MLAH. These strategies aim to minimize the conversion of face-to-face sessions to virtual, ensure fair distribution of teaching responsibilities, and maintain a balanced allocation of face-to-face and virtual classes throughout the academic year. The paper demonstrates the effectiveness of these algorithms in producing high-quality solutions comparable to those generated by CPLEX, with significantly reduced computational complexity. | https://dx.doi.org/10.12785/ijcds/1501124 | Timetabling, Scheduling, Optimization , Blended learning | Mohamed Jarraya (Saudi Electronic University, Saudi Arabia) |
| 125 | 1570994783 | Butterfly Image Identification Using Multilevel Thresholding Segmentation and Convolutional Neural Network Classification with Alexnet Architecture | Lepidoptera is the name for the broad group of butterflies. The ecology depends heavily on butterflies, thus it is problematic that so little is known about their many kinds. Understanding butterflies is a crucial part of education since they are a natural occurrence and may be used as teaching tools. A total of 419 butterfly photos were utilized in the data. The dataset is first input, and then it undergoes preprocessing steps like segmentation, scaling, and RGB to grayscale conversion. CNN with AlexNet architecture is used to classify the preprocessed dataset's output. The outcomes of the classification stage of the Alexnet architecture are Flatten, Danse, and ReLu (Convolution, Batch Normalization, Max_Pooling). The output data is assessed following the completion of the Alexnet CNN training process. The data's ultimate classification is based on species. High-accuracy picture classification can be achieved using the model without segmentation, however, this cannot be achieved with multilevel threshold segmentation. According to the test findings, the multilevel threshold segmentation model only attains 62% accuracy, but the segmentation-free model gets 83% accuracy. The test results demonstrate that combining AlexNet architecture with multilevel thresholding segmentation resulted in a classification model that is less accurate in identifying different species of butterflies. By comparing these test results, it is possible to draw the conclusion that the multilevel threshold segmentation model performs less well at information classification than the model without segmentation. | http://dx.doi.org/10.12785/ijcds/1501125 | Convolutional Neural Network; Butterfly; Segmentation; Multilevel; Alexnet | Abdul Fadlil and Ainin Maftukhah (Universitas Ahmad Dahlan, Indonesia); Sunardi Sunardi (UAD, Indonesia); Tole Sutikno (Universitas Ahmad Dahlan & Universiti Teknologi Malaysia, Indonesia) |
125 papers.