		FeedBack \| Contact Us \| Links \| Site Map

Welcome to Journal of Theoretical and Applied Information Technology

Submit Paper / Call for Papers

Journal receives papers in continuous flow and we will consider articles from a wide range of Information Technology disciplines encompassing the most basic research to the most innovative technologies. Please submit your papers electronically to our submission system at http://jatit.org/submit_paper.php in an MSWord, Pdf or compatible format so that they may be evaluated for publication in the upcoming issue. This journal uses a blinded review process; please remember to include all your personal identifiable information in the manuscript before submitting it for review, we will edit the necessary information at our side. Submissions to JATIT should be full research / review papers (properly indicated below main title).

Journal of Theoretical and Applied Information Technology
October 2024 | Vol. 102 No.20

Title:	FACIAL EXPRESSION ANALYSIS FOR ACADEMIC ENGAGEMENT MONITORING WITH KRIGING GEOMETRY AND REGION PROPOSAL NETWORK DEEP TRANSFER LEARNING
Author:	NOORA C.T, P.TAMIL SELVAN
Abstract:	Over the past few years, contemporary educational movements have made it feasible to apprehend students' facial expressions during erudition to identify learners' poignant status. A growing insistence has been exhibited on utilizing deep learning and the Internet of Things (IoT) to algorithmically identify and elucidate facial changes in various settings owing to technological advancements. Analyzing facial expressions desires to recognize human emotions by exploring visual face information. , This work proposes a method called Kriging Face Geometry and Region Proposal Network-based Deep Transfer Learning (KFG-RPNDTL) to gauge students' level of involvement in the classroom by analyzing their faces and gestures. First, Face Geometry Point Extractor (i.e., face-related feature extraction) is performed to obtain or acquire face-related feature points in an extensive manner. The proposed KFG-RPNDTL method is based on the Euclidean Polar Coordinate Distance Circumplex two-dimensional models of emotions, and it uses the kriging predictor of Best Linear Unbiased Predictors, which minimizes the prediction error. The classification problem related to academic engagement monitoring by analyzing facial expressions has been formulated and solved. The relationship of different emotions is evaluated by plotting different emotions as the points on the plane. The objective is to arrive at an estimate of picture emotion on the plane by kriging and determining which emotion is identified as the closest one. Six basic emotions (Boredom, Confusion, Drowsiness, Engaged, Frustration and Neutral) have been selected. The proposed KFG-RPNDTL method recommends that the Deep Transfer Learning basis of a multimodal scheme precisely find out student academic rendezvous. Experimental research attained an accuracy of 98.85% as well as demonstrated which student academic engagement method is considerably better than existing methods in various metrics through minimal error rate.
Keywords:	Internet Of Things, Facial Expression, Deep Transfer Learning, Kriging Face Geometry, Region Proposal Network, Student Engagement Management
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	COMMUNICATION VIOLENCE EXPERIENCE TOWARDS MILLENNIALS: A PHENOMENOLOGICAL STUDY OF THE ONLINE LOAN APPLICATION CONSUMERS
Author:	SHINTA MURSIATI ANNISA, MEILANI DHAMAYANTI, LA MANI
Abstract:	The growth of financial technology is a real trend. In Indonesia financial technology users are increasing rapidly, one type of financial technology that is growing rapidly is fintech peer to peer lending using mobile application which usually called an fintech lending or online loan application. This study explores and analyses the experiences of millennials regarding communication violence experienced when using online loan application using a qualitative research method with a phenomenological approach and in-depth interview data collection methods with 5 participants. The results of this study indicate that the process of applying for loan through online loan application is very easy with its simplicity and practicality. However, behind the convenience offered, online loan application can also cause psychological trauma for its users which has impact on their daily activities, such as feeling stressed, afraid, anxious, worried, panicked, and other negative feelings due to verbal violence done by debt collectors if the borrower is late in making debt settlement. The implications of this study recommend related institutions to improve the provision of public services in the form of educational services and increasing literacy to the public regarding the use of online loan application. This study also shows that online loan application can have good and bad impacts on their users, so make sure to use online loan application wisely.
Keywords:	Financial Technology, Online Loan Application, Communication Violence, Verbal Violence, Phenomenology, Millennials.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	HEART DISEASE PREDICTION USING CASE BASED REASONING (CBR)
Author:	MOHAIMINUL ISLAM BHUIYAN, CHAN HUE WAH, NUR SHAZWANI KAMARUDIN, NUR HAFIEZA ISMAIL, AHMAD FAKHRI AB NASIR
Abstract:	This study provides an overview of heart disease prediction using an intelligent system. Predicting disease accurately is crucial in the medical field, but traditional methods relying solely on a doctor's experience often lack precision. To address this limitation, intelligent systems are applied as an alternative to traditional approaches. While various intelligent system methods exist, this study focuses on three: Fuzzy Logic, Neural Networks, and Case-Based Reasoning (CBR). A comparison of these techniques in terms of accuracy was conducted, and ultimately, Case-Based Reasoning (CBR) was selected for heart disease prediction. In the prediction phase, the heart disease dataset underwent data pre-processing to clean the data and data splitting to separate it into training and testing sets. The chosen intelligent system was then employed to predict heart disease outcomes based on the processed data. The experiment concluded with Case-Based Reasoning (CBR) achieving a notable accuracy rate of 97.95% in predicting heart disease. The findings also revealed that the probability of heart disease was 57.76% for males and 42.24% for females. Further analysis from related studies suggests that factors such as smoking and alcohol consumption are significant contributors to heart disease, particularly among males.
Keywords:	Case Base Reasoning (CBR), Machine Learning, Heart Disease
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	MILLENIALS IN JAKARTA AND THEIR EXPERIENCES WITH CYBERBULLYING ON INSTAGRAM
Author:	CANTIKA NUR OKTAVIANA, LA MANI, MUHAMMAD AZHAR LAZUARDI, NI PUTU SARI DARMAYANTI
Abstract:	This qualitative research aims to describe and understand the phenomenon of cyberbullying on Instagram within the context of millennial generation’s experiences in Jakarta. Using a phenomenological approach and descriptive qualitative methods, the study explores how millennials experience cyberbullying, the types of content that are frequently targeted, and changes in their self-perception after experiencing cyberbullying. The findings show that cyberbullying on Instagram takes many forms, including offensive comments (flaming), repeated harassment, stalking, spreading false information (denigration), pretending to be someone else (impersonation), and deceptive threats (trickery). Many perpetrators use fake or anonymous accounts to carry out their attacks. The most common targets of cyberbullying are personal posts, such as selfies and updates about achievements. This often leads to a decrease in victims' self-confidence and negative changes in how they view themselves. The study emphasizes the need for effective policies and interventions to address cyberbullying and highlights the importance of educating users about safe and responsible social media use. By providing insights into the challenges faced by millennials in Jakarta, the research aims to support the development of strategies to better protect individuals and reduce the harmful effects of cyberbullying.
Keywords:	Cyberbullying, Instagram, Millennial, Phenomenology, Social media, Good health and Well- Being
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	IMPROVING DEMAND FORECASTING ACCURACY WITH MACHINE LEARNING MODELS: CASE STUDY OF AN INDONESIAN FMCG COMPANY
Author:	FRIZA ARMEN, RIYANTO JAYADI
Abstract:	The FMCG industry in Indonesia is experiencing rapid growth but faces challenges in accurate demand forecasting. This can lead to operational inefficiencies and unnecessary costs. This study aims to improve demand forecasting accuracy by applying various machine learning models and identifying the best model for specific product categories. The study uses historical sales data from an FMCG company in Indonesia to evaluate the performance of seven machine learning models: Simple Moving Average (SMA), Weighted Moving Average (WMA), Exponential Moving Average (EMA), ARIMA, Linear Regression (LR), Artificial Neural Network (ANN), and Long Short-Term Memory (LSTM). The results indicate that the Exponential Moving Average (EMA) model consistently outperforms others across all product categories. Specifically, EMA achieves MAPE values as low as 0.22% in Instant Food and 0.24% in Beverages. This study recommends that FMCG companies in Indonesia use Exponential Moving Averages to improve demand forecasting accuracy. Additionally, the study contributes valuable insights to industry knowledge by providing new perspectives on effective forecasting techniques.
Keywords:	FMCG, demand forecasting, machine learning, Exponential Moving Average, Indonesia
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	MITIGATING DATA SPOOFING RISKS IN NEAR FIELD COMMUNICATION (NFC) READ/WRITE MODE: AN INVESTIGATION INTO ACCESS CONTROL VULNERABILITIES AND POST-COMPROMISE RECOVERY STRATEGY
Author:	PUTERI SHARIZA MEGAT KHALID, NOR FAZLIDA MOHD SANI
Abstract:	The rapid growth of Near Field Communication (NFC) technology has facilitated its widespread adoption in everyday activities, particularly in cashless mobile payments and access control among urban dwellers. However, this convenience is accompanied by significant security risks, including data spoofing, relay attacks, and unauthorized data access. Despite advancements in NFC technology, a critical gap persists in securing communications, particularly at the end-user level, where awareness and preventive measures are insufficient. This research aims to address those gaps by focusing on end-user vulnerabilities and providing tailored solutions through the development of an NFC End-User specific security policy. The study's primary contribution lies in the development of an information security policy tailored specifically for NFC End-Users. This policy serves as a comprehensive guideline aimed at enhancing the security posture of individuals who utilize NFC-enabled devices. Unlike previous studies that primarily examine technical countermeasures, this work emphasizes the human factors by assessing user awareness and NFC secure practices. By developing an NFC security policy specifically for end users, the study aims to bridge the gap between technological safeguards and NFC End User behavior.
Keywords:	Near Field Communication (NFC) End-User Security Awareness, Information Governance, Personal Data Security, Information Security Policy.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	AN INTELLIGENT APPROACH TO CREDIT CARD FRAUD DETECTION USING RANDOM FOREST
Author:	Bhukya Dharma, Dr.D. Latha
Abstract:	Nowadays digitalization gaining popularity because of seamless, easy and convenience use of ecommerce. A credit card which remains a very widespread compensation method is accepted online & offline that provides cashless transactions. Credit card fraud is a critical issue for financial institutions and their customers. Credit card fraud is one of the most important threats that affect people as well as companies across the world, particularly with the growing volume of financial transactions using credit cards every day. Machine Learning algorithms have been applied for identifying fraudulent transactions efficiently. This paper presents, An Intelligent Approach to Credit Card Fraud Detection Using Random Forest (RF). The major issues in fraud detection on credit card transaction data are that they are huge and they exhibit huge imbalance levels. E-Commerce Sales Dataset is obtained from the Kaggle. In the dataset 85275 are the genuine transactions and 117 are fraud transactions. The results of the described model are based on Accuracy, Sensitivity, Specificity, and F1-score. Described model achieves Accuracy as 97.1%, Precision as 95.7%, Sensitivity as 95%, Specificity as 95.9%, and F1-Score as 97.5%. The investigational outcomes absolutely show the effectiveness of described model.
Keywords:	Credit card fraud, Machine Learning, Random Forest, Accuracy, Sensitivity, Specificity, and F1-score
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	TRUST PRIORITY BASED CLUSTERING MODEL FOR MALICIOUS ATTACKS DETECTION WITH SECURE DATA TRANSMISSION IN SMART GRIDS
Author:	CHADALAVADA NAGA PRIYANKA, NANDHAKUMAR RAMACHANDRAN
Abstract:	The use of Smart Grids (SG) and smart meters is becoming widespread in several nations. Hackers with even a rudimentary understanding of computers can compromise smart meters and conduct cyber attacks. Government security and network operators are at risk in this cyberspace. SG companies should create defensive and preventative measures to lessen the impact of electricity theft on their bottom lines. Cyber assaults can compromise cyber-physical systems. In order to identify a cyber attack on the smart grid, numerous methods have been developed. Among the best security measures, weighted trust-based models are recommended. SGs use clustering model to group smart meters for monitoring them. There can be no trust unless the sensors work as intended, if they can communicate with one another, and if the nodes' servers are reliable. How the nodes have communicated in the past also has a role. This research proposes a smart grid sensor network security technique that is based on trust weights in a clustering model. The total trust of nodes by adding up their direct and indirect trust is performed in the smart grid that in a cluster. The presence of numerous bidirectional communication devices connecting consumers to the grid makes smart grid networks particularly vulnerable to network attacks. Malicious assaults can compromise the Smart Grid Network's backbone infrastructure, which consists of information and communication technologies. For the uninterrupted and effective supply of energy and to generate an accurate bill, it is vital to detect the assault and work on it. A large number of compromised grid communication devices or nodes send a flood of false data or requests to the smart grid network, which can disrupt smart meters, data servers, and the state estimator. As a result, end-user services could be affected. When it comes to protecting the network from malicious attacks, a malicious node detection model is proposed. The innovative model detects and eliminates malicious nodes from the network after successfully differentiating between physical and cyber intrusions. Data transmission in a smart grid is accomplished by wireless technologies. There are a variety of network attacks that could compromise SGs. When it comes to protecting massive communication networks from hostile network attacks, trust models are a key component. To avoid malicious actions in the SGs, this research proposes a Trust Priority based Clustering model to detect and avoid Malicious Attacks (TPbCMA) for secure data transmission and increasing the quality of service levels in smart grids. The proposed model efficiently detects the attacks in the smart grids to maintain quality of service levels. The proposed model achieved 98.7% accuracy in Node Clustering and 99.1% accuracy in Malicious Action Detection. The proposed model when contrasted with the traditional models performs superior than traditional models.
Keywords:	Smart Grid, Smart Meter, Cyber Attacks, Cyber Security, Malicious Attacks, Direct Trust, Indirect Trust, Clustering.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	ENHANCED HISTOGRAM INTEGRATED MORPHOLOGICAL IMAGE QUALITY ENHANCEMENT MODEL USING DERMOSCOPY IMAGES WITH EDGE BASED SEGMENTATION
Author:	V GOPIKRISHNA, K THINAKARAN
Abstract:	Melanoma is a major type of skin cancer, and its accurate and prompt identification is becoming more and more important every day. The development of state-of-the-art models and computer vision methods has made analysis much simpler. Effective skin cancer segmentation and classification methods necessitate well-defined lesions isolated from their surroundings. One typical method for detecting edges is to employ a two-dimensional filter that is trained to rely on big gradients to identify changes in pixel intensity in a scene. The operator is subsequently employed to convolve the picture. Edge detectors accumulate many images and apply a local image processing method to identify sudden changes in an intensity function. There have been a plethora of new proposals for pre-processing skin lesions that aim to assist segmentation algorithms in producing good results. More and more people are losing their lives to melanoma each year. Stage I melanoma diagnosis, on the other hand, are associated with better survival chances. The process of melanoma segmentation is quite laborious since it must take into account both the top and bottom of the tumor. We apply a new approach to improving and segmenting melanoma images. Low contrast in dermoscopy images of the skin is often caused by lighting conditions that vary. Because dermoscopy images of melanoma have low contrast, the lesion tends to blend in with the surrounding skin. In addition, the low contrast makes it difficult to make out a number of visual elements. Because of this, there has to be a way to make dermoscopy pictures more detailed and contrasty. To mitigate the effects of low contrast and improve image quality, a morphological method is proposed in this research. A localized set of both light and dark features can be retrieved from a image using image reconstruction. By removing the dark elements and adding the nearby bright ones, the image quality can be improved. This research presents a Enhanced Histogram Integrated Morphological Image Quality Enhancement Model using Dermoscopy Images with Edge based Segmentation (EHIMIQE-DIES). The proposed model performs feature extraction from the quality images. The proposed model is compared with the traditional methods and the results represent that the proposed model performance is high in image quality enhancement and in segmentation.
Keywords:	Enhanced Histogram, Morphological, Dermoscopy Images, Segmentation, Edge Detection, Image Quality, Feature Extraction.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	MACHINE LEARNING FOR STORY POINT ESTIMATION: DO LARGE LANGUAGE MODELS OUTPERFORM TRADITIONAL METHODS?
Author:	Levi Alexander, Riyanto Jayadi
Abstract:	Our research investigates the performance of machine learning models, particularly Large Language Models (LLMs), in automating story point estimation for Agile software development. Traditional estimation methods relying on human judgment can introduce subjectivity and errors. Recent advances in deep learning and LLMs offer potential improvements in accuracy and consistency, especially in handling complex language tasks. We compare traditional machine learning models such as Random Forest, SVM, and Linear SVM with LLMs like BERT and GPT-2, focusing on both within-project and cross-project story point estimation. While traditional models frequently outperform LLMs in project-specific tasks, LLMs show competitive performance in handling more complex and diverse datasets. Our proposed general model, trained on combined datasets, demonstrates competitive results in structured cross-project estimation scenarios, narrowing the performance gap compared to previous models like Deep-SE and GPT2SP. However, project-specific models still outperform the general model in most cases. Our research highlights the trade-offs between model complexity and performance, showing that traditional models are often more efficient and accurate in structured datasets, whereas LLMs excel in tasks requiring deep language understanding. Further refinement of general models could enhance their applicability across diverse projects.
Keywords:	Deep Learning, Large Language Model, Machine Learning, Software Effort Estimation, Story Point Estimation
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	DYNAMIC ADAPTIVE WITH ANT LION OPTIMIZATION FOR AUTISM DISORDER FACIAL IMAGE CLASSIFICATION
Author:	SUJATHA HANUMANTHARAYAPPA, MANJULA RUDRAGOUDA BHARAMAGOUDRA
Abstract:	Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder that affects social interaction and communication with others. However, similarity among facial images of individuals with autism impacts the classification accuracy and difficult for the model to effectively learn the images. In this research, a proposed Dynamic Adaptive Boundary Adjustment with Ant Lion Optimization (DALO) and Softmax loss to enhance performance of Convolutional Neural Network (CNN) for classification. DALO effectively balances exploration and exploitation, optimizing the search of hyper parameter space in SCNN. The SCNN technique enhances loss function by using softmax to address overfitting. The softmax loss function is used by optimization algorithm to update model weights during training. The Autistic Children dataset (ACD) and ASD Dataset and pre-processing to ensure all pixel values contribute equally during learning process. Feature extraction using MobileNetV2 which utilizes depth-wise while maintaining model capacity. The proposed DALO-SCNN method achieves a precision of 0.982, recall of 0.98 and f1-score of 0.93 and accuracy 97% on ACD dataset. The SCNN method achieves better accuracy of 93.12%, precision of 92.56% of precision, 92.01% of recall and 92.25% of f1-score on ASD dataset, when compared to the existing methods such as CNN and ResNet 50 techniques.
Keywords:	Ant Lion Optimization, Autistic Children Facial Dataset, Autism Spectrum Disorder, Convolutional Neural Network, MobileNetV2
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	ADVANCED MACHINE LEARNING TECHNIQUES FOR REAL-TIME FRAUD DETECTION AND PREVENTION
Author:	NARASIMHA SWAMY BIYYAPU, SURESH BABU CHANDOLU, SHOBANA GORINTLA, NARASIMHA RAO TIRUMALASETTI, ANURADHA CHOKKA, S PHANI PRAVEEN
Abstract:	This Exploration researches the utilization of models like Logistic Regression (LR), Linear Discriminant Analysis (LDA), K-nearest Neighbors (KNN), Characterization and Regression Tree (Truck), Naive Bayes (NB), Support Vector Machine (SVM), Irregular Woodland (RF), XGBoost, and LightGBM for real-time fraud detection utilizing a charge card exchange dataset. Head Part Analysis (PCA) was utilized to guarantee information protection and Engineered Minority Oversampling Strategy (Destroyed) was utilized to settle class irregularity in the dataset, which included 284,807 exchanges and 492 fraud occurrences. Utilizing Irregular Timberland to survey highlight pertinence, 27 significant qualities were found. AUC, F1-score, Recall, Precision, KS, and PRAUC were among the performance indicators used to assess the models. Random Forest outperformed the rest in terms of accuracy (99.99%), recall (99.99%), precision (99.98%), and F1-score (99.99%), proving its superiority in separating transactions that are fraudulent from those that are not. The results imply that RF is a very successful model for on-the-spot fraud detection.
Keywords:	Machine Learning, Techniques, Real-Time, Fraud Detection, Prevention
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	THE IMPACT OF ONLINE LEARNING FATIGUE ON STUDENTS' CONTINUOUS USE OF ONLINE LEARNING
Author:	AYMAN ALARABIAT
Abstract:	The shift from face-to-face learning to online learning (OL) becomes more challenging due to the emergence of so-called OL fatigue. Simply, OL fatigue is the degree to which students feel a sense of overload from immersion in the use of OL. Despite the importance of OL fatigue in education, it can be observed that there have not been many studies on the role of OL fatigue and the potential impact on students' continuous use of OL. This study addresses a significant gap in OL literature by examining the impact of OL fatigue on students' continuous use of OL. The analysis of 233 respondents using partial least squares structural equation modeling reveals that four aspects of OL fatigue have negative significant impact on students’ continuous use of OL. The OL fatigue aspects include the burden of the online course (e.g., poorly designed OL courses and weak interaction with instructors and fellow students), psychological challenges (e.g., feelings of isolation, loneliness, stress, and anxiety), lack of sensory requirements during OL (e.g., inability to see and hear educators perfectly and the absence of the physical campus sensation), and the home learning environment. These factors explain around 52.6% of the variance in students’ continuous use of OL, which could be considered relatively substantial in studies that seek to predict human behavioral intentions, as is the case in our study. The current study results highlight the necessity of addressing OL fatigue to ensure long-term OL usage. The study complements and extends the understanding of factors influencing students' continuous use of OL by considering the impact of OL fatigue that has not been previously widely examined. The findings and recommendations provide higher education policymakers with a clearer understanding of students' OL fatigue, which should be reflected in current and future OL policies and regulations.
Keywords:	Online learning, E-Learning, Online learning Fatigue, Continuous use, Post-Adoption
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	PERFORMANCE ANALYSIS OF LT-BCH CODE FOR WIRELESS BODY AREA NETWORK
Author:	LYDIA SARI, SYAH ALAM, INDRA SURJATI
Abstract:	Wireless Body Area Network (WBAN) is a technology developed for various applications, especially in the health sector. WBAN is a network of sensors, actuators and transmitters worn by humans for continuous monitoring of physiological data, which will be subsequently transmitted to a data processor for health diagnosis purposes. Data reliability is of utmost importance in WBAN as erroneously received data can result in misdiagnosis by healthcare professionals. In addition to data reliability, the implementation of a suitable channel code for WBAN must also take energy efficiency into consideration, as WBAN devices typically have limited energy. This research presents the analysis of energy consumption for the Luby Transform (LT) code with Bose-Chauduri-Hocquenghem (BCH) as the outer code, where the energy consumption is compared to SNR per bit and transmission distance. The simulation results shows that the LT-BCH code demonstrates optimal energy consumption performance for WBAN applications. The use of LT-BCH codes with high error correction capabilities, namely LT-BCH(127,64,21) and (255,71,59), is suitable for high transmission distances and poor channel conditions. Results show that code rate has minimal impacts under good channel conditions, which is signified by the converging energy consumption requirements needed for various codes in high SNR regions. The results suggest that the LT-BCH code is an energy-efficient solution for WBAN, particularly in challenging transmission environments.
Keywords:	LT Code, BCH, WBAN, Rayleigh, fading channel
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	5 SCIENTIFIC PROBLEMS FOR IMMERSIVE INTERACTIONS DESIGN WITHIN VIRTUAL REALITY APPLICATIONS DEVELOPMENT PIPELINE
Author:	NURUL AIMAN ABDUL RAHIM, MOHD ADILI NORASIKIN, ZULISMAN MAKSOM
Abstract:	Virtual reality (VR) has emerged as a transformative technology with multifaceted applications across gaming, education, and industry. It is also align with Sustainable Development Goal (SDG) in Industry, Innovation, and Infrastructure, by fostering innovation and technological advancement. However, the inherent complexity of VR application development presents challenges, particularly in immersive interaction design within the development pipeline. In this review, we emphasize the importance of addressing five scientific problems pertaining to interaction design within the VR application development pipeline. By shedding light on these challenges, we aim to stimulate discussion, propose solutions, and ultimately contribute to the advancement of VR techniques and applications, with a particular focus on immersive interaction design. We endeavor to support the realization of SDGs by harnessing VR technology to drive innovation, enhance accessibility, and promote sustainable development across various sectors.
Keywords:	virtual reality, state of the art, immersive interactions design
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	CRYPTOGRAPHIC SECURE DATA SCIENCE MODEL TO PREVENT CYBER SECURITY USING MACHINE LEARNING MODEL
Author:	Dr. MAZHARUNNISA, AMARENDRA REDDY PANYALA, BADDEPAKA PRASAD, I. SANDHYA, Dr. U. SRILAKSHMI, REKHA GANGULA
Abstract:	Cybersecurity is the practice of protecting systems, networks, and data from digital attacks, theft, and damage. It involves implementing measures and technologies to safeguard information confidentiality, integrity, and availability. Machine learning (ML) is revolutionizing the field of cybersecurity by providing advanced tools and techniques to detect, analyze, and respond to cyber threats more effectively and efficiently. This paper proposed Ethereum Hashing Hyperbolic Cryptography (EHHC) for cyber security in the data science model. The proposed EHHC model comprises the Hyperbolic Curve Cryptography (HCC) model for data science security. With the integration of the HCC model in the Ethereum blockchain hashing is performed for data science data security. The proposed EHHC model is deployed in the Ethereum blockchain for data security for cyber security. The cyber threats are estimated and classified with the machine learning model for the classification of attacks using the CICIDS, UNSW-NB15 and KDD datasets. Through the incorporation of the EHHC model cyber threats are classified and detected for the different simulation environments. The results demonstrated that the proposed EHHC model achieves a higher classification accuracy of 96.1% with a minimal computation time of ~12% than the conventional cryptographic techniques. The results expressed a higher classification for cyber threat detection and classification in the data science environment.
Keywords:	Cyberattack, Data Science, Cryptography, Classification, Machine Learning, Security
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	THE METAHEURISTIC OF HYBRID EVOLUTIONARY WITH BLACK-HOLE ALGORITHM FOR COMBINATORIAL PRODUCT LINE TESTING
Author:	RABATUL ADUNI SULAIMAN, NUREZAYANA ZAINAL, MAZIDAH MAT REJAB, NORHAMREEZA ABDUL HAMID, DAYANG NA JAWAWI, WAN NOOR HAMIZA WAN ALI
Abstract:	Software testing is very challenging due to the demand of product quality. This process is related to an optimization procedure that neglects the cost and effectiveness measures of products. The existing method lacks search process, causing it unable to efficiently find optimal solution. The objective of this study is to propose and evaluate an experimental hybrid technique of Black-Hole (BH) with Binary Particle Swarm Algorithm (BPSO) called BH-BPSO using t-way combinatorial testing for test case generation in SPL. This study proposes the BH-BPSO which is based on modification and integration of BPSO and BH algorithm in the searching process. Evaluation of the proposed work is implemented based on four different sizes of SPL case studies. The result shows that the proposed BH-BPSO is comparatively efficient for test case generation. BH-BPSO managed to outperform existing methods based on total execution time, size of test suite, pairwise coverage and test case redundancy measure for large size case studies. It is concluded that this research shows the feasibility of the proposed approach in the SPL testing. This approach achieved improvement in terms of parallel metaheuristic measure for large size of case studies.
Keywords:	Software Testing, Metaheuristic Algorithm, Optimization Problem, Soft Computing, Software Product Line.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	LOCALIZATION UPGRADING THROUGH DEUCE ADAPTATION WITH MUTABLE AMBIT PREMISED LOCALIZATION ALGORITHM IN WSN
Author:	SUCHETA PANDA, SUSHREE BIBHUPRADA B. PRIYADARSHINI, PRABHAT KUMAR SAHU
Abstract:	In modern era, significant obstacles have been encountered in case of earlier methods for attaining optimization for localization due to indiscriminately individualized developments in Wireless Sensor Network (WSN). In this connection, traditional optimization algorithms have lower computational effectiveness and they fail to converge towards the optimal global state, thereby, overlooking the ranging errors and localization geometry while negatively affecting precision and efficiency pertaining to localization in WSN. The current paper throws light on a new-fangled approach called Extemporaneity Bat Optimization Technique (EBOT) based on Deuce Adaptation, through two significant changes. Thus, a unique approach that improves the bat optimization method's exploratory while conjointly exploiting its characteristics. Further, EBOT adaptation 1 improves global search capabilities leading to better exploration, whereas EBOT adaptation 2 employs an enhanced local search method to promote exploitation. The proffered method also presents a Polarity Metamorphosis Strategy (PMS), which improves crossover and mutation operations thereby boosting the population heterogeneity as well as exploration capacities. Additionally, the strategy recognizes the importance of range errors and localization geometry in the context of positioning procedure. Again, the Mutable Ambit Premised Localization (MAPL) Algorithm is presented to élite primary nodes during triangulation based on an unpretentious assessment to enrich localization geometry. We The method enhances accuracy by considering localization geometry and range errors when estimating the final positioning with proficient adaptability.
Keywords:	APIT, BOA, CT, EBOT, FP-MPP-APIT, LE, MAPL, PMS, RMSE, WSN.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	ENHANCED ECG SIGNAL CLASSIFICATION USING HYBRID CNN-TRANSFORMER MODELS WITH TUNING TECHNIQUES AND GENETIC ALGORITHM OPTIMIZATION
Author:	G VAMSI KRISHNA, SHRAVAN KUMAR AVULA, V VIJAYA KUMAR RAJU, T V HYMA LAKSHMI, PRAVEEN TUMULURU, TATA BALAJI, N.JAYA
Abstract:	Electrocardiogram (ECG) signal classification plays a vital role in detecting cardiovascular diseases, particularly arrhythmias. This study explores two advanced approaches, i.e., the Time-Series Transformer Architecture and the Hybrid CNN-Transformer Model. Several domain-specific enhancements are introduced, including custom positional encoding, dynamic attention mechanisms, and cross-attention layers tailored for ECG signal classification. The Time-Series Transformer initially achieves an accuracy of 96.8%, which is improved to 97.4% through modifications. The Hybrid CNN-Transformer model demonstrates superior performance, reaching 97.8% accuracy initially and improving to 98.2% with modifications. Finally, the best-performing Hybrid CNN-Transformer model is optimized using Genetic Algorithms (GA), achieving a classification accuracy of 98.6%. The novelty of this work lies in the application of transformer models with targeted architectural enhancements and GA-based optimization to achieve state-of-the-art accuracy in ECG signal classification.
Keywords:	ECG Signal Classification, Transformer neural networks, Tuning Techniques, and Optimizers.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	DETERMINATION OF AUGMENTATION METHOD IN CONVOLUTION NEURAL NETWORK USING FOX OPTIMIZATION FOR FISH IMAGE CLASSIFICATION
Author:	FIRMAN WAHYUDI, MOCH. ARIEF SOELEMAN, RICARDUS ANGGI PRAMUNENDAR, PULUNG NURTANTIO ANDONO
Abstract:	The automatic identification of fish species plays a crucial role in various fields such as conservation biology, fisheries management, and biological research. Convolutional Neural Network (CNN) methods have become an effective solution for automating this process from digital images. However, achieving high accuracy requires careful consideration of factors such as the amount and quality of data, image processing methods, feature extraction techniques, classification algorithms, and optimization methods. This study addresses these challenges by proposing a CNN model optimized using the FOX algorithm to select the best augmentation method. The results show that selecting the appropriate augmentation techniques, such as Kmeans Color Quantization, Horizontal Flip, Voronoi, Elastic Transformation, and Contrast Normalization, can significantly improve fish species recognition accuracy up to 98.75% during training. The proposed model also demonstrated strong generalization capability, with a validation accuracy of 96.90%, indicating minimal overfitting. Although it requires intensive training time, this approach proves highly effective for applications demanding high accuracy and good generalization, thus enhancing the understanding and management of marine ecosystems in support of sustainable fishing practices.
Keywords:	Fish Image Classification, Convolutional Neural Network, Augmentation, Feature Selection, FOX Optimization Algorithm
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	ANALYSIS OF FACTORS INFLUENCING CONTINUANCE INTENTION TOWARDS DIGITAL BANK APPLICATIONS
Author:	JEREMY NATHANAEL JURNAWAN , TANTY OKTAVIA
Abstract:	The banking industry is undergoing a new era, an era where a significant portion of its activities is not bound by human resources and physical offices (office-less). Numerous digital banking applications representing various banks can be found. One such example is Bank Jago, which has become the most popular digital bank for the Indonesian people. However, looking at the ratings of each application on the Google Play Store in November 2022, Bank Jago, which ranked first in the survey, had the lowest rating. Therefore, this study will examine and identify factors that that could possibility enhance user intention to continue using digital banking application mediated by user satisfaction. The study took sample from a population of digital banking user in Indonesia and received 259 respondents. This study use a modified framework based on TAM & UTAUT framework with the involvement of the user experience variable as a mediating variable, with other variables added such as feature, security and trust by drawing reference from Google Playstore review to understand what users thought about the app. Structural Equation Modeling was used as analysis method with SmarttPLS 3.0 software. The result indicates that social influence is the only variable that does not have a significant impact on user satisfaction. Meanwhile, features, perceived ease of use, perceived usefulness, security and trust have a positive and significant influence on user satisfaction. Similarly, user satisfaction have a positive and significant influence on users' interest in continuing to use digital banking
Keywords:	Digital Banking, Continuance Intention, SmartPLS, Structural Equation Modeling (SEM)
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	FACTORS AFFECTING CRYPTOCURRENCY ADOPTION INTENTION AMONG INDIVIDUALS
Author:	RABEA ABDULRAHMAN RAWEH, MASLINDA MOHD NADZIR, HUDA HAJI IBRAHIM
Abstract:	Cryptocurrency is a modern sort of virtual currency that operates through blockchain technology and whose purpose is to be employed as a means of exchange. It is currently attracting the attention of academic and non-academic researchers as an alternative digital currency. The rise of cryptocurrency has recently gained a massive increase in cryptocurrency markets all around the globe. However, insufficient attention has been paid to the unveiling of determinants driving cryptocurrency adoption. Thus, the study aimed to fill the gap in the current literature by investigating factors that influence the adoption of cryptocurrency among individuals. The research used a survey questionnaire to gather data from a sample of 270 respondents. Therefore, the collected data was analyzed using structural equation modeling (SEM) and basic descriptive statistics. Furthermore, the results indicated that facilitating conditions, social influence, awareness, and security significantly affect cryptocurrency adoption intention. This study is critical for analyzing and gaining insights into individuals’ primary motives for cryptocurrency adoption, which will help in formulating a regulatory framework.
Keywords:	Cryptocurrency, Adoption Intention, Facilitating conditions, Social Influence, Awareness, Security, Trust.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	RELIABLE POWER-OPTIMISED TOKEN-PASSING ACCESS METHOD COMMUNICATION FOR MAC
Author:	DR. REKHA GANGULA, SREENIVAS PRATAPAGIRI, DR. C MADAN KUMAR, DR.L. MOHAN, DR. VENKATESWARLU.B, DR.A. MANJULA
Abstract:	Reliable Power-Optimized Token-Passing Access Method Communication for MAC (Media Access Control) refers to a networking protocol designed to efficiently manage data transmission in a network, particularly in scenarios where power consumption is a critical concern. This method employs token-passing, where a token circulates among nodes to regulate their access to the network. In this paper proposed mechanism operates by assigning a token for exclusive channel access, coupled with continuous retransmission requests from nodes based on data age. This approach effectively reduces collisions and offers automatic retransmission opportunities to nodes experiencing prolonged transmission failures. Crucially, the token holding time (THT) parameter governs bandwidth allocation per node in the token-ring network, requiring careful calibration to prevent deadline misses. Additionally, the target token rotation time (TTRT) dictates both token circulation speed and network utilization, necessitating meticulous selection to ensure optimal performance. Through extensive simulations, it is demonstrated that our proposed method outperforms existing approaches, achieving a 30% reduction in collision rates and a 20% improvement in successful beacon transmissions. By dynamically adjusting parameters such as token holding time (THT) and target token rotation time (TTRT), our method optimally allocates bandwidth and token circulation speed, ensuring efficient network utilization while minimizing deadline misses. Furthermore, our power optimization strategy, employing clock gating buffers, yields a notable 15% reduction in overall power consumption without sacrificing network performance.
Keywords:	Medium Access Control, Adaptive MAC, Multi-Token Based Collision Free Data Transmission, Token Holding Time, Target Token Rotation Time, Clock Gating.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text

Title:	OBJECT DETECTION AT RAILWAY LEVEL CROSSING TO IMPROVE PUBLIC SAFETY
Author:	TOMMY WISNU WARDHANA, S.Kom., RIYANTO JAYADI, S.Kom.
Abstract:	The most widely used mode of transportation is the train. The growth of Indonesian Railways has greatly benefited the populace in a number of ways, most notably in terms of operations and services. But there hasn't been a corresponding improvement in safety, particularly at railroad level crossings. The paper purpose of this research is to offer ways to improve safety at railway level crossings. The community ultimately benefits from this, as there are fewer accidents at train level crossings. Resources used by making advantage of the CCTV cameras that are stationed at various railroad level crossings. The datasets were created by merging Visdrone datasets, which were supplied by Ultralytics, the company that makes YOLOv8, with additional unique datasets for certain data that are not included in the package. As far as we are aware, this is the first study on object detection at Indonesian level crossings. Although there were some references to comparable research conducted in other nations, none of them made use of YOLOv8, which is now the greatest detection technique The methodology involves making observations and interviews with the division in charge of CCTV surveillance at level crossings. Python programming is done via AnacondaPrompt's Command Line Interface (CLI) tools, while LabelImg is used to provide annotations for custom datasets. The study's findings demonstrate that object identification at railroad crossings is capable of accurately and precisely detecting objects for all object classes, with a high precision level of 98.7%.
Keywords:	Object Detection, Train, Level Crossing, YOLOv8.
Source:	Journal of Theoretical and Applied Information Technology 31^st October 2024 -- Vol. 102. No. 20-- 2024
Full Text