		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
March 2024 | Vol. 102 No.6

Title:	SEMANTIC QUERY EXPANSION METHOD FOR DIGITAL RESOURCE OBJECTS RETRIEVAL
Author:	WAFA’ ZA’AL ALMA’AITAH1 ABDULLAH ZAWAWI TALIB, ALAA OBEIDAT, MOHD AZAM OSMAN, FATIMA N. AL-ASWADI, RAMI S. ALKHAWALDEH
Abstract:	Digital resource objects (DRO) consider as one of the most useful resources for storing humanity's collected knowledge. Many organizations are now aiming to make this data available to individuals. The query provided to DROs by the non-expert user, on the other hand, is usually a brief and frequently confusing expression of his desire. In DROs, it is not enough to explicitly explain what the user requires. The reason for coming up with short user query is that the users usually have limited knowledge and terminologies of the specific domain area. The formative terms can be missing inside the user’s query, leading to poor coverage of relevant documents. To cover the difference between the query of user and DROs, the semantic query expansion method (SQE) is proposed to improve the efficiency of DRO retrieval by enhancing the quality level of candidate terms to be inserted semantically to the entire query terms to enhance performance of DRO retrieval. The proposed SQE method comprises three steps namely query terms definition, candidate terms generation and the proposed correlation algorithm. The aim of the correlation algorithm is to extract the semantic terms to extend the query with related terms only. Results from experiment on CHiC2013 and ECHiC2013_EDE collections show that the proposed method can significantly outperform previous methods specifically in DROs.
Keywords:	Query expansion method, Information retrieval, Digital resource objects, Semantic query
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	ADVANCING DIABETIC FOOT ULCER DETECTION BASED ON RESNET AND GAN INTEGRATION
Author:	AHMED MOSTAFA EL-KADY, MOHAMED M. ABBASSY, HEBA HAMDY ALI, FARID ALI MOUSSA
Abstract:	Diabetes, characterized by the body's inability to effectively regulate sugar levels due to insulin complications, leads to various serious health issues. Among these, Diabetic Foot Ulcer stands out as a critical yet often ignored consequence. This condition, if not addressed in time, can result in severe outcomes including amputations, posing a substantial burden on both individuals and healthcare systems, particularly in areas where medical care is costly. Addressing this pressing issue, our research focused intensively on the analysis of medical images, with the goal of enhancing the accuracy of Diabetic Foot Ulcer diagnosis. We assessed two different models: the renowned ResNet50 model and hybrid model that fuses ResNet50 with Generative Adversarial Networks. The findings were noteworthy; the ResNet50 demonstrated commendable performance, achieving an average accuracy and precision of 0.76, and an F1-Score of 0.75. However, the hybrid model surpassed these metrics, registering an average accuracy of 0.84, precision of 0.85, and an F1-Score of 0.84. This research contributes to the evolving landscape of medical image analysis, offering a promising avenue for more precise and effective DFU diagnosis in clinical settings. The marked advancement in diagnostic precision afforded by the hybrid model suggests a significant stride forward in effectively managing and treating DFU.
Keywords:	Diabetic Foot Ulcers, Deep learning, DFU, ResNet50, Generative Adversarial Networks, GAN.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	CHARTING NEW FRONTIERS: ASSESSING INFORMATION TECHNOLOGY'S ROLE IN THE EVOLUTION OF IMAGE ENHANCEMENT – A BIBLIOMETRIC APPROACH
Author:	RAJIMOL A, JUBY GEORGE, JEENA JOSEPH, JOBIN JOSE, SNEHA M KURIAKOSE, BEENAMOLE T
Abstract:	Image enhancement techniques hold a crucial role within the realm of image processing. The primary aim of these methods is to enhance the visual appeal of images, making them more suitable for subsequent processing or human interpretation. This research investigates the bibliometric analysis of image enhancement methods, providing a comprehensive perspective of the research landscape in this domain using VOSviewer and biblioshiny. Through a comprehensive review of a substantial body of publications, we include the most prominent authors, institutions, and countries driving advancements in the field of image processing. Additionally, the paper highlights the key challenges and recent developments in image enhancement. This study underscores the collaborative networks among researchers and the extensive array of investigations related to image enhancement. The insights from this study offer practitioners, decision-makers, and academics valuable information that guides them toward influential areas and potential strategies for enhancing images in the future. Various methodologies for image enhancement across different domains are explored, which is crucial for image analysis and recognition. This study uniquely underscores the transformative impact of Information Technology in refining image enhancement methods, contributing to both computational efficiency and enhanced interpretability. The revelation of untapped collaborative networks and innovative methodologies, as presented in this research, marks a significant leap in harnessing Information Technology for future breakthroughs in image processing.
Keywords:	Image Enhancement Method, Bibliometric Analysis, VOSviewer, Biblioshiny.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	PROGNOSIS OF CARDIOVASCULAR DISEASE USING PRINCIPAL COMPONENT ANALYSIS AND SUPPORT VECTOR MACHINE CLASSIFICATION ALGORITHMS IN THE R STUDIO ENVIRONMENT
Author:	RAHBRE ISLAM, SAFDAR TANWEER, TABREZ NAFIS, IMRAN HUSSAIN, SHAHAZAD NIWAZI QURASHI
Abstract:	Cardiovascular disease is a significant cause of death throughout the globe. Early detection of this lethal disease can be important to avoid future losses. The underlined issues can be unraveled using patients’ medical history and machine learning algorithms and can predict heart disease status before it gets in worse condition. The predictive ability of ML algorithms, particularly SVM, is promising for cardiovascular diseases. This study also presents machine learning approaches for predicting heart diseases, using data on major health factors from patients. The Principal Component Analysis (PCA) and Support Vector Machine (SVM) have been applied to comprehend and classify patient data. The main aim of this study is to predict heart-related conditions well in advance to avoid any fatality. Complex data can be simplified using PCA while Support Vector Machine helps to assess predictions. The combination of these methods is applied in the R studio environment to assess heart health accurately and efficiently. Data preprocessing and feature selection steps were done before building the models. The accuracy of SVM with and without Principal Component Analysis (PCA) is 90.49% and 84.88% respectively where SVM with PCA outperformed.
Keywords:	Heart Disease, SVM, PCA, R Language, Machine Learning
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	ENERGY OPTIMIZATION APPROACH BASED MACHINE LEARNING ON LINEAR REFLECTOR SYSTEMS
Author:	MOHD ARIF, MALABIKA ADAK, ANJUM ARA AHMAD, CHHAYA NAYAK4, IBRAHIM AQEEL, SHADAB ALAM, DYUTI BANERJEE, AYASHA SIDDIQUA
Abstract:	Solar energy is a renewable and cost-effective energy source that holds great promise for meeting global energy demand. This capability is exemplified in solar-powered cooling systems, which have gained popularity in recent years. Addressing the environmental issues associated with fossil fuel consumption, this study investigates the application of linear Fresnel reflectors (LFRs) in solar-based refrigeration systems A novel approach uses equipment -Uses learning models to optimize LFR output, so that grid-based energy -Increased supply chain efficiency. The study uses Simulink simulations to evaluate the performance of the model, with a particular focus on its energy efficiency compared to other methods. This study addresses the complex challenges posed by sustainable energy and highlights the need for sustainable solutions. Studies focusing on LFRs contribute to the growing knowledge of solar energy efficiency. The inclusion of machine learning techniques demonstrates innovation in improving the performance of solar collectors and offers potential improvements in solving energy consumption problems Simulations in Simulink for testing robust areas, indicate that the proposed machine learning LFR outperforms other methods in terms of energy efficiency. The study highlights the importance of switching to renewable energy sources, especially in cooling systems. The use of linear Fresnel reflectors, together with advances in machine learning, represent a promising approach to energy efficiency. These issues contribute to the ongoing discourse on sustainable energy practices, offering compelling solutions for reducing environmental footprints.
Keywords:	Energy Consumption; Cooling Systems; Machine learning; Linear Fresnel reflectors; Energy Efficiency
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	TRANSFORMATIVE PEDAGOGIES: A BIBLIOMETRIC JOURNEY THROUGH ADAPTIVE LEARNING SYSTEMS
Author:	JOBIN JOSE, ALICE JOSEPH, PRATHEESH ABRAHAM, ROSHNA VARGHESE, BEENAMOLE T, SONY MARY VARGHESE, SUBY ELIZABETH OOMMEN
Abstract:	As a major shift in education technologies, Adaptive Learning Systems (ALS) use artificial intelligence and similar technologies, adapting the lessons to the needs of individual students. Emphasizing transformative pedagogy and teaching strategies that transform the learners' cognitive and interactive patterns, this study presents a comprehensive bibliometric analysis of ASL. Contrary to conventional teaching methods, ALS alters dramatically the way students think and interact with their environment. This research has utilized an all-inclusive bibliometric analysis to analyze the evolution, trends, and themes in ALS by using an extensive set of data from the Web of Science (WoS) and Scopus. The primary objective of Bibliometric analysis is to map the development of ALS in teaching and learning while marking the important trends, models, and thematic priorities. The relevance of this research lies in its comprehensive analysis of the Adaptive Learning Systems (ALS) field through bibliometric methods, offering critical insights into the trends, key contributors, and thematic developments over time. The systematic evaluation enables the appraisal of the impact created by major contributors like authors, organizations, journals, etc. The study also examines, using the advanced data collection technique, influential articles, and publications that enormously contributed to shaping ALS. Similarly, it does the rating effectively upon evaluating the mutual relationships among important terms, concepts, and factors through co-references and co-occurrences. It highlights the increasing scholarly output and identifies key contributors and influential works, underscoring the growing recognition of ALS's importance due to technological advancements. The study's findings on global research contributions, thematic analyses, and collaboration networks offer new insights into the field's dynamics, setting a foundation for future research directions. To visually represent bibliometric data, web analytic tools are used, explaining intricate relationships and thematic clusters. Identifying the unexplored areas and discussing the practical implications of ASL development, research, and analysis of combined data taken from WoS and Scopus provides a unique perspective. Consequently, researchers, educators, policymakers, etc., get valuable insights that enable advancing and understanding the area. This bibliometric analysis will undoubtedly guide future research in the area of transformative pedagogy as it is the most sought-after method in understanding the scholarly landscape of ALS
Keywords:	Adaptive Learning System, Bibliometric Analysis, R-Studio, Biblioshiny
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	AN INTEGRATED DEEP LEARNING BASED ENHANCED GREY WOLF OPTIMIZATION FOR LUNG CANCER PREDICTION
Author:	T. DIVYA, Dr. J. VIJI GRIPSY
Abstract:	Lung cancer is an extremely harmful disease that represents the leading cause of death among both males and females within the nation. The survival spans for lung cancer patients within the 10%-20% range are limited to a duration of five years. Nevertheless, in the event that lung cancer is identified in its early stages and promptly treated, there is potential for a reduction in death rates. When lung cancer is identified at an early stage during the screening procedure, the clinical response to treatment may exhibit variability and provide very favourable outcomes. The implementation of a dependable and automated system might greatly facilitate the early identification of lung cancer, even in remote regions. This research presents a unique technique called Integrated Deep Learning-based Enhanced Grey Wolf Optimization for lung cancer prediction (IDL-EGWO). In order to address the issue of instability and convergence accuracy that occurs when using the Grey Wolf Optimizer (GWO) as a meta-heuristic algorithm with a robust capacity for optimum search, A weighted average GWO algorithm is suggested as a way to try to fix the problems with the GWO, such as the fact that it can get stuck in local optima and has a slow convergence rate in later stages. This technique incorporates an Artificial Neural Network (ANN) during the training phase. The research included a range of performance criteria, including precision, recall, f-measure, accuracy, execution time, and root mean squared error. According to the experiment, the IDL-EGWO algorithm demonstrated a higher accuracy rate of 97% compared to the previous methods.
Keywords:	Lung Cancer Prediction, Optimization, Deep Learning, GWO, ANN, MLP.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	EMPOWERING SENTIMENT ANALYSIS OF COURSERA COURSE REVIEWS WITH SOPHISTICATED ARTIFICIAL BEE COLONY-INSPIRED DEEP Q-NETWORKS (SABC-DQN)
Author:	J SAHITHA BANU, G PREETHI
Abstract:	This paper presents SABC-DQN (Sophisticated Artificial Bee Colony Inspired Deep Q-Networks), a novel approach to enhance the sentiment analysis of Coursera course reviews. Sentiment analysis is crucial for understanding learner opinions, but existing methods struggle to capture the nuanced sentiments expressed in textual data accurately. SABC-DQN combines the intelligent exploration capabilities of Artificial Bee Colony (ABC) optimization algorithms with the power of Deep Q-Networks (DQN). The ABC optimization algorithms mimic honey bees’ efficient foraging behaviour, enabling effective solution space exploration. DQN, a reinforcement learning technique, utilizes a deep neural network to learn and approximate the optimal policy for sentiment classification. The SABC-DQN approach operates through a multi-step process. Initially, the ABC optimization algorithm guides the exploration of the solution space, identifying optimal features that capture sentiment-related information. These features are then employed to train the DQN, leveraging the representation learning capabilities of the deep neural network to predict sentiment labels accurately. Experimental evaluations conducted on a dataset of Coursera course reviews demonstrate the efficacy of SABC-DQN in enhancing sentiment analysis. The proposed approach outperforms existing methods, achieving superior accuracy, precision, recall, and F1 score. SABC-DQN exhibits robustness when faced with variations in review length, domain-specific jargon, and grammatical errors. SABC-DQN introduces a novel solution for empowering sentiment analysis of Coursera course reviews. By integrating sophisticated Artificial Bee Colony optimization algorithms with Deep Q-Networks, SABC-DQN provides an advanced mechanism to capture nuanced sentiments expressed in textual data. The proposed approach has the potential to significantly improve sentiment analysis accuracy, facilitating a deeper understanding of learner perspectives in online educational platforms.
Keywords:	Artificial Bee Colony, Coursera, Deep Q-Networks, SABC-DQN, Sentiment Analysis, Textual Data Analysis
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	VALIDATION OF SECURE E-VOTING SYSTEM BASED BLOCKCHAIN IMMUTABILITY: THE JORDANIAN PARLIAMENTARY ELECTIONS
Author:	KHALID ALTARAWNEH, AMER OSHOUSH, IBRAHIM ALTARAWNI, MOHAMMED AMIN ALMAIAH, TAYSEER ALKDOUR, ABDALWALI LUTFI, MAHMOUD AL-RAWAD, AND RAMI SHEHAB
Abstract:	This paper aims at providing a valid E-voting framework schema based on BC technology that serves the election process in the Jordan context. It investigates the validity and feasibility of the proposed model using a focus group study to restructure the implementation process as an initial step before the researchers develop the final software product that aligns with Jordanian parliament election law. The study initially proposes a framework for an e-voting framework based on blockchain technology and then evaluated it through a focus group. The proposed framework schema fulfills the standards of adopting E-voting by stressing a set of principles related to consistency, integrity, and identity verification. The final conceptual framework schema was developed considering the validation results recommended by the experts' evaluation to become a valid model that can move forward in following up on its implementation.
Keywords:	Security, E-Voting System, Blockchain, Jordan Parliamentarians, Elections
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	TRANSFORMATIVE LEARNING MODEL WITH DIGITAL FABRICATION LABORATORY TO ENHANCE INNOVATION COMPETENCY AND CREATIVE PRODUCT
Author:	SUNTI SOPAPRADIT, PRACHYANUN NILSOOK, PANITA WANNAPIROON
Abstract:	This research’s objectives were 1. To develop system of transformative learning with digital fabrication laboratory to enhance innovation competency and creative product. 2. To study the results of the demonstration system. The study included 2 process: 1) the development system of transformative learning with digital fabrication laboratory to enhance innovation competency and creative product, and 2) the result of the demonstration system. The use of transformative learning with digital fabrication laboratory was demonstrated to students in Southeast Bangkok College who registered for the ‘Industry Innovation Laboratory’ subject in their first semester of B.E. 2565. The findings revealed that the scores with regard to 1) the curriculum, 2) the context quality, 3) the quality of the lecturing media, and 4) the quality of the learning system were rated as being at the highest level. Moreover, after studying the subject, the experimental group had greater innovation competency scores than they exhibited prior to the learning process. In addition, the experimental group had higher scores with regard to innovation competency and creative product after studying than the control group and the criteria. A system of transformative learning with digital fabrication laboratory to enhance innovation competency and creative product is a new teaching model that can develop students' skills in the 21st century. Therefore, it should be applied to other subjects in this field by only subjects that are practiced so that students can develop their own skills.
Keywords:	Transformative Learning, Digital Fabrication Laboratory, Innovation Competency, Creative Product
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	A NEW INTELLIGENT SUGENO WEIGHTED FUZZY BASED SOPHISTICATED COMPUTATIONAL MODEL TO ANALYSE LEADERSHIP ELEGANCE ON ESTABLISHMENT
Author:	K. KRISHNA RAO, DR. KVB GANESH
Abstract:	Technology is the new normal now a days, without that noting is perfect and frequent, in this article we emphasises on a New intelligent Sugeno Weighted Fuzzy based Sophisticated Computational Model to Analyse Leadership Elegance on Establishment of any organisation. The purpose of this research is to examine the effects of different leadership styles on productivity in the power sector, with a focus on the APGENCO (Dr. Narla Tata Rao Thermal Power Station) in Ibrahimpatnam. The presented model employs the Sugeno Weighted Fuzzy Model, a sophisticated computational approach, the study enhances its analytical precision. For the purpose of study the researcher has focused on five fuzzy aspects as leadership styles such as autocratic, democratic, Lassiez-faire, Transformational and transactional leadership styles. The fuzzy logic-based methodology is particularly suitable for capturing the inherWent uncertainties and complexities present in human behavior and organizational dynamics. For the purpose of data analysis, the researcher used a descriptive correlation method. For the purpose of sampling the researcher used a stratified random sampling method to gather the data. To test the hypothesis, the researchers used the correlation method using SPSS software. The results showed that leadership style has a postive impact on employee performance.
Keywords:	Weighted Fuzzy, Computational Model, Leadership Styles, performance, SPSS software.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	A MATHEMATICAL FRAMEWORK TO DECODE ATTENTION FROM CORTICAL RESPONSES FOR HEARING-IMPAIRED LISTENERS AND BCI APPLICATIONS
Author:	ANUDEEP PEDDI, DR. T. VENKATA RAMANA
Abstract:	The science keeps advancing rapidly, new research topics keep emerging daily, and one such field is neuroscience. New computer models are being proposed that mimic human visual and auditory systems, with the former being the central area of focus. Humans are very good at focusing their attention on a required sound. This is not possible for people with hearing impairment because hearing aids amplify all the incoming signals. Our objective is to try and model the auditory system of humans, specifically on the topic of auditory attention. Our ears are always active and are fed with a large variety of sounds at each moment. We aim to model when our attention is grabbed by a particular sound amongst a large cacophony of sounds. If this is implemented on a hardware system, people with hearing issues can focus only on required sounds. This can be developed by using the concept of temporal response functions (TRFs), which show the linear relation between audio and EEG signals. We proposed a new mathematical framework to overcome the current challenges to predict the sound envelope. This obtained envelope is compared with the audio input given while the EEG data was recorded, using the concept of correlation. The correlation coefficients obtained for different values of regularization parameters are discussed. The proposed mathematical technique gave a better result compared to the existing state-of-the-art techniques.
Keywords:	Mathematical Modeling, Regression, Cocktail Party Problem, Auditory Scene Analysis, Auditory Attention Detection, Hearing Impairment solutions, Regression.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	SENTIMENT MINING OF CUSTOMER REVIEWS FROM E-COMMERCE WEBSITES
Author:	KUNDETI NAGA PRASANTHI, MVP CHANDRASEKHARARAO, SUDHASREE CHEKURI, SESHU BABU P
Abstract:	Over the past few decades e-commerce has increased manifolds. The e-commerce websites ask their customers t0 share their views ab0ut the pr0ducts they have purchased. Therefore, milli0ns 0f reviews are accumulated for the products in e-commerce websites. Customers view the reviews of the product before they purchase the product. If a product has more positive reviews then that will result in more customers buying the product. So, classifying the vast collections of reviews into different categories is the need of the hour. This paper describes one of such mechanisms for classifying reviews using text mining and extracting the sentiment of the review. The proposed mechanism involves extracting product reviews from e-commerce websites, identifying the terms that represent the sentiment of products, highlighting the positive terms that are more frequent and specifying the frequencies of different sentiment defining terms in the product reviews.
Keywords:	Classification, Amazon Reviews, Sentiment Analysis, Wordnet, Positive Word Cloud
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	A NOVEL METHOD FOR INDIAN NUMBER PLATE DETECTION AND RECOGNITION USING EFFICIENT NET
Author:	MR.U. GANESH NAIDU, DR. R.THIRUVENGATANADHAN, DR. S. NARAYANA, DR. P. DHANALAKSHMI
Abstract:	Number Plate Recognition (NPR) is essential in supporting the government in properly managing vehicles as the number of private vehicles increases significantly throughout the world. However, different number plate types or slight variations to the number plate format can disrupt existing NPR systems because they fail to detect the number plate. Additionally, the NPR system is very responsive to environmental factors. To properly address these issues, this research introduces an innovative deep learning-based NPR system. A robust NPR system that integrates three pre-processing algorithms including super-resolution, low-light enhancement, and defogging. And the number plate is effectively-recognized by using these algorithms in a range of environments, it is one of this paper's research achievements. Then the number plate is successfully segmented by applying contours through border following and filtering the contours based on spatial localization and character dimensions. Finally, the EfficientNet algorithm is used for character recognition after de-skewing and region of interest filtering. The ImageAI library is used by the proposed deep learning model to enhance training. Images of Indian number plates are utilized to evaluate the model's performance. The accuracy of 99. 2% is achieved for number plate detection and an accuracy of 98.78% is achieved for character recognition. The extensive performance is achieved by the proposed method compared to previous methods. The implementation is performed under the python platform.
Keywords:	Number Plate; Super-Resolution; Spatial Localization; Defogging: Segmentation: Character Recognition.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	OBJECT DETECTION OF CHILI USING CONVOLUTIONAL NEURAL NETWORK YOLOV7
Author:	RICHARD SALIM, AHMAD NURUL FAJAR
Abstract:	In Indonesia, the production of red curly chili faces challenges in stabilizing market prices, leading to a growing dependence on chili imports to maintain stability. Import figures surged by 237.07% in early 2023, rising from 1.24 million kilograms in January 2022 to 4.18 million kilograms. This reliance on imports is primarily due to the rigid distribution system of chili peppers, which closely follows farmers' harvest schedules. Consequently, inconsistent chili availability and uncertain quality result from prolonged regional distribution times, impacting market prices. The sorting process is crucial in determining prices for all participants within the Indonesian chili supply chain. Unfortunately, the current manual sorting process is plagued with shortcomings, negatively affecting the efficiency of the entire chili supply chain. Therefore, it is crucial to develop innovative strategies to aid supply chain participants in chili cultivation and boost chili sales by automating the sorting process. In this research initiative, our team proposes a solution involving the development of the YOLOv7 model for automatic detection and classification of high-quality red curly chili. Our approach included collecting image data, rigorous data preprocessing, and hyperparameter optimization. The YOLOv7 model demonstrated commendable performance, achieving an impressive overall grade Mean Average Precision (mAP) of 0.977. Additionally, it exhibited noteworthy average precisions (AP) with scores of 0.996 for grade A, 0.947 for grade B, 0.951 for grade C, and 0.996 for grades D and E.
Keywords:	Deep Learning, Object Detection, Object Classification, YOLOv7.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	STORAGE STRUCTURES IN THE ERA OF BIG DATA: FROM DATA WAREHOUSE TO LAKEHOUSE
Author:	MOHSSINE BENTAIB, ABDELAZIZ ETTAOUFIK, ABDERRAHIM TRAGHA, MOHAMED AZZOUAZI
Abstract:	The amount of data that is available to enterprises today comes from many different sources, including social networks, sensors, and IoT devices. In order to discover trends, draw conclusions, produce projections, and make informed decisions, this enormous amount of data needs to be stored across a variety of platforms for processing and analysis. The capacity of conventional EDs is surpassed by the quantity and quality of data that is being collected. To accomplish this, businesses with current data warehouses must pick a storage architecture with enough storage and processing power for this kind of data. They must choose one of the following options: The data warehouse can either (i) develop into a big data warehouse, (ii) be replaced by a data lake, or (iii) be combined with a data lake to create a data LakeHouse. In this article, we aim to find the best choice for the storage of varied and voluminous data. To do this, we examine the big data warehousing literature. After doing a comparison of the various architectures put forth, we draw a conclusion outlining the optimum storage practice.
Keywords:	Data Warehouse, Big Data, Big Data Warehouse, Data Lake, Data Lakehouse
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	MULTI-OBJECTIVE METHOD COMBINATION ANALYSIS OPTIMIZATION ON THE BASIS ON RATION ANALYSIS (MOORA) AND BEST FIRST SEARCH ALGORITHM IN THE SELECTION OF OBSTETRICIAN AND GYNECOLOGY PRACTICES
Author:	HANDRIZAL, FAUZAN NURAHMADI, ABRAHAM ALDIO SIANIPAR
Abstract:	The health of pregnant women must be given great attention to assist in the delivery process and prevent maternal death. Based on the results of data from the Ministry of Health, currently, the maternal mortality rate in Indonesia is relatively high, as much as 32% of maternal deaths in Indonesia occur due to bleeding and another 26% occur due to hypertension which can cause seizures and poisoning. Routine check-ups with an obstetrician are one of the right steps to maintain the health of mothers and children during pregnancy. Therefore, this study tries to develop an application that can help pregnant women to find recommendations for obstetricians and gynecologists. By using the Multi-Objective Optimization based on Ratio Analysis (MOORA) method is to find out where obstetricians and gynecologists practice recommendations and the Best First Search (BFS) method is used to determine the optimal sequence of selecting practices based on the results of the MOORA method. This study is on patient satisfaction, service quality, accessibility, and cost efficiency. Based on the results of the user acceptance test, the results show that this system can meet demand and can be used properly.
Keywords:	Multi-Objective Optimization Based On Ratio Analysis (MOORA), Best First Search Algorithm, Obstetrics And Gynecology Practices, Practice Selection.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	ENHANCING MALWARE DETECTION EFFICACY: A COMPARATIVE ANALYSIS OF ENDPOINT SECURITY AND APPLICATION WHITELISTING
Author:	MOHAMMED ALTHAMIR, ABDULLAH ALABDULHAY, KHALED RIAD, ABDULLAH ALBUALI
Abstract:	Endpoint security solutions are increasingly critical in light of the continual expansion of cyber threats, particularly malware, and the growing complexity of threat actors. Leveraging innovative techniques such as AI-based malware detection is necessary to counteract the increasing sophistication of malware. Additionally, alternative solutions like application whitelisting have been developed to protect users from malware infections by only permitting whitelisted applications to run on a host's real operating system. Safeguarding endpoints serves as a primary defense against cyberattacks through comprehensive security protocols, allowing organizations to better navigate the intricate digital environment fraught with potential risks. In this study, we evaluate four pivotal endpoint security solutions: Network Detection and Response (NDR), Endpoint Detection and Response (EDR), application whitelisting, and antivirus software with a specific emphasis on their ability to detect and handle malware threats.The findings of this study provide valuable insights into the effectiveness of application whitelisting compared to antivirus, EDR, and NDR endpoint solutions.
Keywords:	EDR, NDR, Antivirus, Application Whitelisting, Malware Detection, and Endpoint.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	SYSTEMATIC LITERATURE REVIEW ON THE USE OF MACHINE LEARNING IN ONLINE LEARNING IN THE CONTEXT OF SKILL ACHIEVEMENT
Author:	YEFTA CHRISTIAN, YUN-HUOY CHOO, NOOR FAZILLA ABD YUSOF
Abstract:	Online education has experienced significant progress, with machine learning playing a crucial role in improving the outcomes of skills acquisition. In this examination of scholarly literature, we analyze research focused on the use of machine learning in online education, aiming to achieve specific skills. The results of this comprehensive literature review reveal a wide range of machine learning applications, including adaptive modeling, personalized content, automated grading, and student progress forecasting. This approach has enhanced the effectiveness of online education by providing a more targeted and personalized learning experience for each student. Additionally, several studies demonstrate notable improvements in skill acquisition, as indicated by various measures. However, challenges arise regarding data privacy, model accuracy, and the need for validating outcomes in real-world contexts. Therefore, this systematic literature study presents a comprehensive overview of the implementation of machine learning in the field of online education, with the ultimate objective of skills acquisition. This investigation offers valuable insights into the potential and limitations associated with the use of machine learning in online education. These insights will greatly benefit educators, researchers, and developers of educational platforms in understanding how this technology can be utilized to enhance students' proficiency in the continually evolving landscape of online learning.
Keywords:	Skill, Online Learning, Education, Machine Learning, Learning Analytic
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	INTERNET OF THINGS (IOT) ADOPTION IN REMOTE AUDIT: A QUANTITATIVE STUDY APPLYING THE TECHNOLOGY ACCEPTANCE MODEL
Author:	NATHANIA PUTRI MAHARANI, CYNTHIA RAHEL SALIM, BAMBANG LEO HANDOKO
Abstract:	Internet of Things (IoT) system has been adopted by many in various industries as it facilitated the completion of people’s work, including audit. IoT enables its users to collect and process enormous amounts of data, while also to monitor and track the data among IoT-based devices in real-time. For that reason, IoT system has been applied to support auditors doing their work remotely. This study is designed to assess users’ acceptance of IoT-based technology, focusing on the field of audit that is done remotely in public accounting firms in Indonesia. This research uses the Technology Acceptance Model (TAM) that measures the adoption of new technologies with perceived ease of use and perceived usefulness to study users’ intentions to adopt a technology, while also adding perceived enjoyment to measure the degree to which the user perceives a usage of a certain system to be enjoyable. This research incorporates these factors as the independent variables to investigate auditors' intention to use IoT-based remote audit processes. Finally, audit firm size is used as moderating variable to see if there are any changes made on perceived ease of use, perceived usefulness, and perceived enjoyment’s effect on the intention to adopt IoT-based remote audit, that was caused by the moderation. The data collected from 100 auditors in public accounting firms located in Indonesia is then used to conduct the hypothesis testing using Partial Least Squares Structural Equation Modeling. The research’s conclusions showed that the adoption of IoT-based remote audit is significantly influenced by perceived enjoyment rather than perceived usefulness or perceived ease of use. Furthermore, the size of the audit firm did not exhibit moderating capabilities on the influence of implementing IoT-based remote audit.
Keywords:	Internet of Things (IoT), Remote Audit, Auditors, Technology Acceptance Model, Intention to Adopt
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	BREAST CANCER IMAGE CLASSIFICATION USING CUSTOM CNN
Author:	KALLA YOGESWARA RAO, KONDA SRINIVASA RAO
Abstract:	Background: Breast cancer remains a primary global health concern, emphasizing the critical need for accurate diagnostic tools. This study focuses on developing a precise method for classifying breast cancer images using a specifically designed Convolutional Neural Network (CNN). The research employs the BreakHis dataset for training and evaluation, comprising high-resolution histopathological images of breast biopsy specimens stained with hematoxylin and eosin. Methods Used: The unique CNN architecture incorporates convolutional layers, max-pooling layers, dropout layers, and batch normalization, tailored to capture intricate patterns distinguishing between benign and cancerous breast tissues. Comprehensive data preprocessing is implemented, involving label extraction from filenames and augmentation techniques to enhance the training set. The training of the CNN model involves using the Adam optimizer, binary cross-entropy loss, and evaluation metrics such as binary accuracy and ROC-AUC. Early halting and learning rate decrease callbacks are integrated into the training process to optimize model performance. Results Achieved: The trained CNN model is assessed on a separate test dataset, and performance metrics, including ROC-AUC, accuracy, and a confusion matrix, are provided. The findings demonstrate that the custom CNN reliably categorizes breast cancer images, suggesting its potential as a valuable tool for automated breast cancer diagnosis. Notably, the study reports a high ROC-AUC value (0.98051) and satisfactory accuracy (0.93285), indicating the effectiveness of the custom CNN for breast cancer histopathology image categorization. Concluding Remarks: This work underscores the significance of tailored CNN architectures in enhancing the precision of breast cancer diagnostics, contributing to the ongoing efforts to leverage machine learning in histopathological image processing. The promising outcomes of the proposed approach set the stage for further advancements in computer-aided diagnostics and medical image analysis. The reported high ROC-AUC value and accuracy affirm the efficiency of the custom CNN, supporting its potential application in real-world breast cancer diagnostic scenarios.
Keywords:	Breast Cancer, Global Health, Convolutional Neural Network (CNN), Diagnostic Instruments, BreakHis dataset.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	A NOVEL MODEL FOR SECURING SEALS USING BLOCKCHAIN AND DIGITAL SIGNATURE BASED ON QUICK RESPONSE CODES
Author:	MAY WEZA, M. M. EL-GAYAR, AHMED ABO ELFETOH, MOHAMMED ELMOGY
Abstract:	In an era where personal identification, academic achievements, and other critical records are increasingly digitized, the integrity of such documents is paramount. Conventional validation methods, such as stamp seal imprints from authoritative bodies, are under siege by sophisticated counterfeiters, leveraging technological advancements to undermine document authenticity. This study introduces a groundbreaking dual-strategy framework to fortify the security of these stamp seals through orchestrated partial and full digitalization. Our multifaceted approach synthesizes an array of authenticity indicators—encompassing stamp seal images, digital signatures, watermarks, and distinct textual elements—within a robust feature extraction and analysis protocol. This protocol is meticulously engineered to validate the integrity of both physical and digital documentation. Central to our framework is integrating a decentralized blockchain platform, which serves as a bastion for the encrypted authenticity data, ensuring a tamper-resistant, transparent, and distributable ledger of the stamp seals without any reliance on intermediaries. Complementing this, we generate a Quick Response (QR) code for each document, which serves as a portal to interface swiftly and securely with the blockchain record. Our comprehensive testing yields an impressive 98% accuracy and security rate in document and stamp seal imprint verification, markedly enhancing retrieval speeds. The culmination of our research is establishing a rapid, secure, and immutable verification system that significantly eclipses traditional centralized methods, heralding a new standard in document security.
Keywords:	Forged Stamp Seal Imprint; Digital Signature; QR Code; Blockchain; Smart Securing Model
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	WHITE HOLE ATTACKER DETECTION IN MOBILE ADHOC NETWORK
Author:	Dr.S.HEMALATHA, SONIA MARIA D'SOUZA, KHADRI SYED FAIZZ AHMAD, M.RAJASEKARAN, PANKAJ RANGAREE, P. SUKANIA, M.POMPAPATHI, ASHOK BEKKANTI
Abstract:	While making communication among the wireless nodes, which relies on without making infrastructure less network are vulnerable to security fall. One of the most affecting vulnerable security falling wireless networks is Mobile Adhoc Network. The most predominant kind of security falls are intruders and attackers whose roles are trying to diminish the internal performance of the Network. Many research works are concentrating to detect and prevent these two factors. This article concentrates on predicting white hole attackers inside the communication or not. White hole attackers is a kind of attacker whose role is to send the many duplicate packets to the neighboring node to increase the load of the neighbor nodes which affect the overall Mobile Adhoc network performance . Many existing research used the latest technique to predict the attackers which are additional overload to the network .To achieve this objective the WatchDog method introduces to monitoring the forwarded time of the every nodes present in the communication a node which make plenty of times forwarded the packet to the many nodes assumes as white hole attackers. The proposed Watchdog Algorithm with Classification Technique was implemented with Network simulator and the simulation results are compared with Machine learning based routing protocol then the compared results are proved the WatchDog based attacker methods performs well is more than 30 % better also the performance factors are excellent in 60%.
Keywords:	MANET, Attackers, White Hole Attackers, WatchDog Technique, Forward time, Threshold Value
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	SECURE AND EFFICIENT DATA SHARING SCHEME FOR MULTI-USER AND MULTI-OWNER SCENARIO IN FEDERATED CLOUD COMPUTING
Author:	Dr. IMTIYAZ KHAN, Dr. A.YASHWANTH REDDY, Dr. MANIZA HIJAB, Dr. KOTARI SRIDEVI, Dr. SYED SHABBEER AHMAD, Dr. D.SHRAVANI
Abstract:	Cloud computing, since its inception, has undergone continuous improvements. Now federated cloud is realized with seamless integration of diversified clouds. In this context, it is essentially and multi-owner and multi-user environment where security to data of data owners is to be given paramount importance. Supporting data sharing with security in place and enabling users to perform keyword search on the encrypted content is indispensable in such environment. The existing schemes suffer from performance issues in complex multi-owner and multi-user scenarios in federated cloud setting. To address this problem, in this paper, we proposed a security scheme that enables efficient data sharing across users. Users are able to access data of multiple data owners by generating trapdoors. Different algorithms are proposed to realize the scheme. With empirical study, it is observed that our scheme is able to support secure and efficient data sharing in federated cloud environment. Our scheme performs better than existing ones in terms of storage overhead and execution time.
Keywords:	Secure Data Sharing, Cloud Computing, Federated Cloud, Aggregate Key Sharing, Cloud Data Security
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	EVALUATING THE EFFECTIVENESS OF QUILLBOT IN IMPROVING STUDENTS' PARAPHRASING SKILLS: TEACHERS’ VOICES
Author:	TAJ MOHAMMAD, ALI ABBAS FALAH ALZUBI, MOHD NAZIM, SOADA IDRIS KHAN
Abstract:	Information technology (IT) has greatly contributed by ushering into a new era in education by facilitating personalized learning, simplifying information access, and empowering teachers to produce multimedia-rich content using AI-mediated digital tools like QuillBot. QuillBot, an AI-mediated cutting-edge educational technology, is employed to facilitate teachers' pedagogical endeavors across academic orientations, including academic writing classes. The most recent literature exemplifies a surge in using AI applications in developing EFL writing skills; however, studies show that teachers’ voices are not heard enough. Therefore, the current study aims to evaluate the effectiveness of QuillBot in improving students' paraphrasing skills from their EFL teachers’ perspective. To achieve the study's objectives, the descriptive survey method was employed. A randomly stratified sample of (40) teachers teaching in the preparatory year responded to a closed-item questionnaire and a semi-structured interview. The results showed that the study sample highly evaluated QuillBot as a capable tool to improve students' paraphrasing skills. In addition, educational qualifications and experiences did not contribute to varying their responses. Finally, the interview revealed some advantages and disadvantages of using QuillBot. In light of the findings, the study recommends that QuillBot should be integrated as an educational resource within the writing curriculum.
Keywords:	Effectiveness, EFL Teachers, QuillBot, Paraphrasing Skills, Preparatory Year Students
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	DEEP LEARNING CENTERED METHODOLOGY FOR AGE GUESSTIMATE OF FACIAL IMAGES
Author:	MR.A.ASHOK BABU, DR.G.SUDHAVANI, DR.P.VENU MADHAV, P.SADHARMASASTA, KURRA UPENDRA CHOWDARY, T.BALAJI, DR.N.JAYA
Abstract:	Since the rise of social platforms and online entertainment, it is now relevant to a growing number of purposes to determine an individual's apparent age from a facial image. Due to its many uses in fields like security, recruiting, validation, and intelligent social robots, it is a crucial task. It is difficult and time-consuming to use facial images to estimate a person's age with reasonable accuracy. Recently, Convolutional Neural Network (CNN) has demonstrated exceptional performance when analyzing images of human faces. The accessibility of datasets for preparing and an expansion in computational power has made profound learning with Convolutional Neural Network a superior strategy for age assessment. In this task , the proposed CNN model requires less preparation information and furthermore keeps a low Mean Absolute Error (MAE).The model ResNet 50 carries out age assessment as a relapse issue. For the preparation stage, two datasets in particular APPA-REAL and UTKFace are utilized and for the testing stage FG-Net dataset is utilized. In spite of having more modest dataset than past works, the MAE is not exactly past works.
Keywords:	MAE, CNN, ResNet50,CNN,UTKF
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	DIABETES DIAGNOSIS AND LEVEL OF CARE FUZZY RULE-BASED MODEL UTILIZING SUPERVISED MACHINE LEARNING FOR CLASSIFICATION AND PREDICTION
Author:	TEH NORANIS MOHD ARIS, AZURALIZA ABU BAKAR, NORMADIAH MAHIDDIN, MASLINA ZOLKEPLI
Abstract:	A reliable medical decision-making is essential to diagnose a disease. This assists medical practitioners to detect a disease at early stage especially diabetes that causes further health complications. The diversity and availability of healthcare datasets supports medical practitioners to use computer applications in the diagnosis process. There are many medical datasets available for research usage but these datasets lacks information that allows decisions to be made accurately, which have a major impact to diagnose a disease. Fuzzy logic has contributed to handle vagueness and uncertainty issues and one of the appropriate models for the development of medical diagnostics. Most computer applications use machine learning and data mining techniques to aid classification and prediction of a disease. Therefore, a fuzzy model based on machine learning and data mining is a vital solution. In this study, ten supervised machine learning algorithms namely the J48, Logistic, NaiveBayes Updateable, RandomTree, BayesNet, AdaBoostM1, Random Forest, Multilayer Perceptron, Bagging and Stacking are applied for a simulated diabetes fuzzy dataset, verified by medical experts. The fuzzy datasets provide adequate information on the type of diabetes diagnosis and level of care related to the type of diabetes diagnosis. All algorithms were compared based on the accuracy, precision, recall, F1-Score, and confusion matrix. Experiment results for diabetes diagnosis dataset indicate 100% accuracy for the eight algorithms except AdaBoostM1 which produced 79.82% accuracy and Stacking 67.89% accuracy. In addition, level of care dataset reveals the highest accuracy of 97.15% for MLP and Bagging algorithms and the lowest accuracy of 91.66% for stacking algorithm. Overall, the proposed fuzzy rule-based diabetes diagnosis and level of care fuzzy model works well with most of the machine learning algorithms tested. Therefore, the proposed fuzzy model is a useful aid in the decision-making process, specifically in the healthcare sector.
Keywords:	Decision-making, Fuzzy, Supervised Machine Learning, Classification, Prediction
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	DETERMINATION OF LFG QUALITY FOR OPTIMIZING PRODUCTION WASTE POWER PLANT USING FUZZY ANALYTICAL HIERARCHY PROCESS
Author:	MUHAMAD HADDIN, AGUS FAHRUDDIN, ARIEF MARWANTO
Abstract:	This study discusses the determination of the production quality of the Waste Power Plant (WPP). The purpose of this research is to calculate and predict a rough estimate of bio gas production, potential gas emissions and potential estimates of electrical energy generated from determining the quality of landfill gas production. The model is determined as WPP with input of waste volume, gas parameters, weather and the resulting output is quality gas production and electrical energy potential. The parameters used are volume of waste; stockpiling area; gas concentration: CH4; CO2; O2; H2S; and weather. With a constant average condition of waste methane decomposes 64% fast, 14% decomposes slowly, and 22% slowly decomposes. Furthermore, Fuzzy Analytical Hierarchy Process (FAHP) is used to establish the priority weights of the LFG criterion in decision support systems and fuzzy logic is implemented to determine the value of LFG production quality. The research object was carried out at the Jatibarang WPP, Semarang, Indonesia. The results showed that the FAHP can provide an output value of gas quality with an accuracy rate of 79%, with the sanitary landfill model producing a maximum potential of 2.6 MW of electrical energy. Meanwhile, gas emissions released into the air in 2021 are 24,780 tons/year of CH4 and 1,425 tons/year of CO2, while the factors that most influence the quality of LFG gas are: methane gas content, carbon dioxide and weather conditions.
Keywords:	LFG Quality, Waste Power Plant, Fuzzy AHP
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	RESEARCH INTELLIGENT PRECISION MARKETING OF INSURANCE BASED ON EXPLAINABLE MACHINE LEARNING: A CASE STUDY OF AN INSURANCE COMPANY
Author:	NOUHAILA EL KOUFI, ABDESSAMAD BELANGOUR
Abstract:	Today, being a marketer is not an easy task, as it requires guiding relevant interactions with customers and driving business success. This is particularly challenging in the realm of traditional marketing. Over the past few years, marketers have observed that they are spending a significant amount of money on advertising their brands or services without any assurance of a response from the customers who receive their direct mail. This lack of knowledge about their audience makes it difficult to identify the interactors from the non-interactors, leaving marketers feeling like they are marketing blindly. They operate without knowing if they are reaching the right audience at the right time, which further complicates the issue and prolongs the process of creating engagements and building an audience for their brands or services. The primary goal of any marketer is to reduce costs and increase revenues. With the widespread digitalization of services and communication technology in different domains, like the insurance sector, online platforms are producing a huge amount of data every day about customer behaviors. Thanks to this source of information, and driven by new challenges in the market, realizing a more accurate and intelligent marketing approach becomes an increasing necessity among researchers and companies. This study presents an intelligent system based on the combination of advanced features engineering approaches and machine learning techniques. The aim of the suggested precision-making system is to assist managers in discerning customer categories based on potential characteristics. Firstly, a comprehensive customer persona was developed by extracting a range of data features, including basic attributes and consumption attributes. Then, we evaluated the effectiveness of various algorithms, such as CatBoost, XGBoost, random forest (RF), k-nearest neighbor (K-NN), nave Bayes (NB), and support vector machine (SVM) methods, for predicting the response of existing customers to the next offer. Various feature selection techniques were employed to determine the most significant features. Furthermore, the performance of the models used was assessed and compared. The results showed that CatBoost had higher accuracy, kappa, precision, Fmeasure and AUC values of 0.871, 0.711, 0.94, 0.822, and 0.85, respectively, outperforming the other models. To illustrate the advantages of our proposed precision-making system, we used a real-world dataset from an American insurance company as a case study.
Keywords:	Precision Marketing, Machine Learning, Features Engineering, Big Data Analysis, Customer Persona, Decision-making System
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	ADVANCING HEART DISEASE DIAGNOSIS AND ECG CLASSIFICATION USING MACHINE LEARNING
Author:	CHEIKH ABDELKADER AHMED TELMOUD, MOUSTAPHA MOHAMED SALECK, MOHAMEDOU CHEIKH TOURAD
Abstract:	Cardiovascular diseases, encompassing diverse heart-related conditions, present a significant global health challenge. Precise and timely diagnosis is crucial for effective management, driving us to explore data science and machine learning for heart disease prognosis and electrocardiogram (ECG) classification. Using three datasets, including a well-established heart dataset, we harnessed various models, such as Decision Tree (DT), Random Forest Classifier (RF), Support Vector Machine (SVM), K- Nearest Neighbor (KNN), and Neural Network (NN), to predict heart diseases. Remarkably, both RF and DT models achieved a remarkable accuracy of 99 %. In Heart Disease Prediction, advanced techniques such as Gradient Boosting (GB) and Logistic Regression (LR) have been employed to elevate precision boundaries. NN achieved 81% precision and was closely trailed by RF at 80%. The standout performer LR achieved an impressive value of 93%, setting a new benchmark. Our efforts were extended to ECG classification using MIT-BIH. By leveraging robust RF, we achieved a remarkable 97% accuracy, highlighting the potential of machine learning in cardiac health. This study con-ducted a comprehensive comparative analysis of supervised learning algorithms, in which RF emerged as the most precise algorithm. These findings aim to catalyze a new era of precision cardiac diagnostics, introducing unparalleled accuracy and efficiency in heart disease prognosis and ECG classification.
Keywords:	Heart Disease Prediction, ECG Classification, Logistic Regression, Random Forest, Gradient Boosting, Cardiac Health.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	OBJECT DETECTION USING CONVOLUTIONAL NEURAL NETWORK YOLOV7 TO DETECT BANANA RIPENESS
Author:	DICKY ANTONY, AHMAD NURUL FAJAR
Abstract:	This study addresses the challenge of ensuring fruit quality in Indonesia, the 8th largest fruit producing country globally. Despite favorable environmental conditions, many harvested fruits fail to meet quality standards due to various factors such as inadequate water content and soil conditions. To tackle this issue, Convolutional Neural Network (CNN) modeling is employed to assess the quality of golden bananas. This study utilizes the YoloV7 model to detect bananas based on skin color, distinguishing between grade A and grade B bananas. The model achieves a mean Average Precision (mAP) of 78.1%, with grade A achieving 99.5% and grade B achieving 56.7% in Average Precision (AP). These findings contribute to enhancing fruit quality assessment methods and offer a potential solution to improve the quality of harvested fruits.
Keywords:	Object Detection, CNN, YoloV7. Deep Learning.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	A GENERIC FRAMEWORK FOR DEVELOPING REGULATORY TECHNOLOGY AND SUPERVISORY TECHNOLOGY
Author:	ADHITYA DWI KRISTANTO, ARRY AKHMAD ARMAN
Abstract:	Regulatory Technology (RegTech) is the use of technology to ensure regulatory compliance. RegTech has been used not only by financial industries but also by non-financial industries. On the other hand, Supervisory Technology (SupTech) is the use of technology to facilitate and enhance supervision activities and processes. RegTech and SupTech are underpinned by popular technologies, such as Big Data Analytics, Artificial Intelligence, Machine Learning, etc. However, it is hard to find RegTech / SupTech discussions in computer science or system information disciplines. Hence, no reference can guide an organization to develop RegTech/SupTech. Therefore, it is necessary to design a framework for developing RegTech/SupTech. The framework is developed using Design Science Research Methodology (DSRM). The proposed framework consists of nine components, i.e., regulation, risk management, environment constraint, functionality, technology, data, people, project strategy & management, and repository. Each of the components suggests recommended output(s) in order to develop a RegTech/SupTech. The proposed framework covers the components of existing frameworks and gives two additional components. Hence, this research contributes to provide an alternative framework for developing RegTech/SupTech.
Keywords:	RegTech, SupTech, Development, Framework, Generic
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	ARTIFICIAL BEE COLONY OPTIMIZATION WITH FEATURE FUSION BASED AGRICULTURAL PLANT DISEASE DETECTION AND CLASSIFICATION MODEL
Author:	A. PAVITHRA, G. KALPANA
Abstract:	Agricultural plant disease detection and classification is a critical task to ensure productivity and health of the crops. Laboratory testing, visual inspection, and the use of technology such as imaging and machine learning (ML) are different methodologies used for the detection and classification of plant disease in agriculture. Visual inspection includes inspecting the appearance of the plant and symptoms like wilting, abnormal growth, and discolouration. The ML algorithm can be used for the classification of various plant diseases and trained to identify patterns in an image of diseased plants. This method could help researchers and farmers to accurately and quickly recognize plant diseases and take necessary measures to control or prevent them. This study develops an Artificial Bee Colony Optimization with Feature Fusion based Agricultural Plant Disease Detection and Classification (ABCFF-PDDC) model. The presented ABCFF-PDDC technique focuses on the detection of plant diseases via computer vision and feature fusion concepts. In the presented ABCFF-PDDC technique, NestNet model is initially used for the background removal process, i.e. segmenting the leaf regions in the image. Next, deep instance segmentation (DIS) is applied for the segmentation of diseased leaf regions. For feature extraction, a fusion based feature extraction comprising EfficientNet and residual network (ResNet101) model with Nadam optimizer is used. Finally, the ABC algorithm with 1D convolutional neural network (1D-CNN) is used. The experimental analysis of the ABCFF-PDDC model on benchmark plant disease dataset reported the betterment of the ABCFF-PDDC technique in terms of different measures.
Keywords:	Agriculture; Feature fusion, Computer vision, Deep instance segmentation, Parameter tuning, Plant disease detection
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	OPTIMIZING SHOPPING EXPERIENCES: BAYESIAN OPTIMIZATION-INSPIRED RANDOM FOREST FOR ENHANCED SENTIMENT ANALYSIS
Author:	R. ANITHA, D. VIMAL KUMAR
Abstract:	This article investigates the optimization of sentiment analysis within online shopping. It begins by outlining the inherent challenges in sentiment analysis, particularly in handling large volumes of unstructured textual data and achieving accurate results amidst noise and context complexities. In response to these challenges, the proposed work introduces the utilization of Bayesian Optimization-inspired Random Forest (BO-RF) as a solution. This innovative approach aims to enhance sentiment analysis accuracy and efficiency by leveraging the combined strengths of Bayesian Optimization and Random Forest algorithms. The mechanism of the proposed work involves leveraging Bayesian Optimization to tune the hyperparameters of the Random Forest model efficiently. This process optimizes the model's performance, improving the accuracy of sentiment analysis tasks. BO-RF excels in feature selection, enabling the extraction of relevant information from text data while filtering out noise. Through comprehensive experiments and evaluations, the results demonstrate the superior performance of BO-RF compared to traditional sentiment analysis methods. The proposed approach achieves higher accuracy rates and greater efficiency in sentiment classification tasks, showcasing its potential to revolutionize sentiment analysis within online shopping.
Keywords:	Sentiment Analysis, Bayesian Optimization, Random Forest, Online Shopping, Accuracy
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	INTEGRATING ADVANCED CONVOLUTIONAL NEURAL NETWORKS AND IOT IN HEALTH MONITORING: A NOVEL APPROACH TO REAL-TIME HEALTH ANOMALY DETECTION AND RISK STRATIFICATION THROUGH MULTI-SENSOR DATA ANALYSIS
Author:	ANTONY PRADEESH D, SUBIRAMANIYAM N P
Abstract:	The problem addressed in this research is a timely one in modern healthcare: the establishment of advanced health monitoring systems that are capable of drawing intelligent inferences and predictions in real time. By bringing together Internet of Things (IoT) technology and Deep Learning techniques, this research proposes a new approach to the realization of remote patient monitoring (RPM). This system is capable of achieving pervasiveness with today’s vital signs — such as blood pressure, heart rate, oxygen saturation level and cerebral blood flow data — and can provide fall detection using radar sensors, as well as air pollution analysis in indoor environments, thus making the system augmentative and pervasive, not just comprehensive. The research demonstrates the highest accuracy in health anomaly detection and health risk stratification when utilizing Advanced Convolutional Neural Networks for Health Anomaly Detection (ACN-HAD) which are developed and Health Risk Stratification Neural Networks (HRS-NN) to realize the potential for both personalized healthcare and predictive analysis, thus advancing the state of the art in smart healthcare technologies. The results presented have significant potential to revolutionize healthcare delivery and its efficiencies, with the prospect for better patient outcomes.
Keywords:	Remote Patient Monitoring, Internet of Things, Deep Learning, Health Anomaly Detection, Predictive Healthcare
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	THE EFFECT OF TECHNOLOGICAL, ENVIRONMENTAL, AND INDIVIDUAL CONTEXTS ON CBEG: THE ROLE OF TRUST AS A MEDIATOR
Author:	WISSAM H. S. ALKHILANI, RODZIAH BINTI ATAN, SALFARINA BINTI ABDULLAH, YUSMADI YAH BT JUSOH
Abstract:	Cloud-based E-Government (CBEG) offers several advantages to governments. However, the adoption rates of CBEG among institutions of public sectors in developing countries are still limited. The E-Government development and participation index in Iraq is lower than the regional and global averages. This research seeks to pinpoint the most influential factors affecting the intention to adopt CBEG from the perspectives of technological, individual and environmental contexts in Iraq. Based on relevant theories, this study proposed that technological, environmental and individual factors affect the CBEG. In addition, trust was proposed as a mediator. Data are collected from 366 decision-makers as the target respondents of this study through five institutions responsible for the E-Government project in Iraq. Data analysis was performed with SMART-PLS 4. The findings showed that technological, individual, and environmental variables had statistically significant effects on the adoption of CBEG. Social influence was insignificant. Further, trust mediated the proposed relationships. More focus on technological, environmental, and individual factors will enhance the adoption of CBEG.
Keywords:	CBEG, TOE, DOI, SET, Trust.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	DISASTER CONTROL SYSTEM FOR LANDSLIDES USING SUGENO FUZZY ALGORITHM
Author:	VERDI YASIN, ZULFIAN AZMI, IFAN JUNAEDI, ANTON ZULKARNAIN SIANIPAR, ITO RIRIS IMMASARI, OKTAVIA MARPAUNG
Abstract:	One of the geological phenomena known as landslides is the movement of rocks or soil of different kinds, such as when boulders fall or heaps of soil collide. Natural disasters like landslides cause a great deal of death and property damage, not just in Indonesia but all over the world. Landslide incidents typically result from a variety of causes. One of them is the idea that rivers and rainwater cause erosion. Science must therefore find a way to stop this landslide disaster before it happens and prevents people from losing their property and suffering fatalities. Thus, by making use of this system for controlling landslides. This study determines whether or not development on a passed hill or mountain is susceptible to landslides. where the Fuzzy Sugeno algorithm's computation yields the expected results of the process on this component research system tool.The first sensors are the MPU-6050, which can detect movement on the ground, the hygrometer sensors, which can detect soil moisture, and the wemos d1 r2, which can send messages to residents via Telegram bots and trigger a buzzer by sounding a relay.
Keywords:	Fuzzy Sugeno; Avalanche; System Control.
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	IMPROVING SECURITY: BLOCKCHAIN BASED IOT SOLUTIONS FOR THE HEALTHCARE
Author:	PAVITHRA P S, DURGADEVI P
Abstract:	Blockchain and the IoT together provide a new paradigm with the potential to transform healthcare as we know it. This study delves into the revolutionary possibilities of blockchain technology as it pertains to healthcare IoT application security and optimization. Data security, integrity, and interoperability are major concerns with healthcare IoT devices since they produce large volumes of personal patient data. A strong answer to these problems is blockchain technology, which uses a distributed and unchangeable record. Integrity of data, audit trails that are easy to see, and safe information exchange among stakeholders are all possible with blockchain protocols in healthcare IoT systems, all while protecting patients' privacy and obtaining their consent. In addition, the article explores practical examples of healthcare IoT systems enabled by blockchain that have been shown to be beneficial. Secure patient data sharing among healthcare providers, tamper-proof administration of medical records, and supply chain authentication of medications are all examples. Also discussed are some of the restrictions and difficulties that come with using blockchain technology in healthcare IoT, including issues with scalability, regulatory compliance, and interoperability. We provide solutions to these problems and hope that they will encourage blockchain technology to proliferate throughout healthcare IoT networks.
Keywords:	IOT, Blockchain, Healthcare, Security, RSA Algorithm
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	INTELLIGENT ALZHEIMER’S DISEASE PREDICTION USING EXPLAINABLE BOOSTING MACHINE
Author:	ARCHANA MENON P, R. GUNASUNDARI
Abstract:	Alzheimer’s Disease (AD), a progressive brain disorder, poses a growing health challenge. Early detection is crucial for providing proper treatment and to prevent its progression. Revolutionary deep learning models used in AD prediction exhibit high performance compared to simpler models while its black box nature makes the model capricious for the clinicians to make decisions. This paper aims to propose a ML model for the accurate detection and prediction of AD in an explainable way. Feature selection techniques are employed to maximize the relevancy of features with the class labels. Among the different glass-box and black-box models inspected to prognosticate AD, Explainable Boosting Machine (EBM) with Chi-square feature selection could generate more accurate and explainable results even in small datasets. The interpretability graph of EBM delivers both global and local explanation for the predicted results and identifies the features responsible for pulling the prediction towards a particular class. EBMs foster trust and transparency in model’s decision-making process. The proposed model obliges the medical practitioners to take better and confident decisions.
Keywords:	Alzheimer’s Disease, Black-box, Explainability, Explainable Boosting Machine, Glass-box, Feature Selection
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	COMPUTER-AIDED SCREENING AND DIAGNOSIS SYSTEM FOR GLAUCOMA CLASSIFICATION USING DEEP LEARNING
Author:	HARI KRISHNA KANAGALA, DR. V.V. JAYARAMA KRISHNAIAH
Abstract:	Vision is a fundamental human sense, facilitated primarily by the eyes. However, the eye is vulnerable to damage, often resulting in vision loss due to injuries or diseases such as glaucoma. While early detection and accurate identification of glaucoma are crucial for preventing vision impairment, existing research has primarily focused on general glaucoma detection rather than differentiating between its various types. Leveraging advancements in imaging technology, this paper explores the utilization of computer vision and image processing techniques for near-instant diagnosis of different types of glaucoma. Specifically, we propose a method utilizing multi-layer perceptron and particle swarm optimization (PSO) to classify different varieties of glaucoma. However, our experimental results indicate accuracies in the proposed mechanism's ability to detect glaucoma subtypes effectively.
Keywords:	Glaucoma, Deep Learning, Multi-Layer Perceptron, Optimization
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text

Title:	A NEW APPROACH FOR DETECTING KINDS OF CHRONIC KIDNEY DISEASES BASED ON DATA MINING APPROACHES
Author:	JAMALALDEEN SALMAN, SARA ELHISHI, SAMIR ABDELRAZEK, HAZEM ELBAKRY
Abstract:	Chronic Kidney Disease (CKD) is a potentially fatal condition that can last a person's whole life and is caused by either kidney cancer or impaired kidney function. It is possible to stop or limit the progression of this chronic condition to the point when dialysis or surgical intervention are the only possibilities for saving the life of a patient. Earlier detection and treatment can reduce the likelihood of this occurring. This study investigates how Machine Learning (ML) techniques can be used to detect different kidney disease kinds. ML algorithms have been a driving force in the detection of abnormalities in various physiological data and are being used successfully in various classification tasks. In this study, we employed six different supervised- machine learning algorithms such as Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), K-Nearest-Neighbor (KNN), XGBoost, and Random Forest (RF) in two different experiments for detecting kidney disease types. Each experiment used a different dataset taken from Kaggle (CKD dataset) and Alliance website (kidney disease genes dataset) for building binary and multi classification kidney disease models. SVC and KNN achieved 99.00%, 99.21% accuracy and recall, respectively for first experiment. While KNN and LR achieved 62.22% and 83.33% accuracy and precision, respectively for second experiment with adequate robustness, and our findings imply that KNN can also be used to detect similar diseases.
Keywords:	Machine Learning, chronic kidney disease, KNN, RF, LR, DT, SVM, XGBoost
Source:	Journal of Theoretical and Applied Information Technology 31^st March 2024 -- Vol. 102. No. 6-- 2024
Full Text