		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
May 2025 | Vol. 103 No.10

Title:	SOL-AUTOCLUST: A SMART ONLINE-LEARNING AUTOMATED CLUSTERING FRAMEWORK
Author:	IBRAHIM GOMAA, HODA M. O. MOKHTAR , NEAMAT EL-TAZI , ALI ZIDANE
Abstract:	The automation of machine learning has predominantly focused on supervised tasks, leaving unsupervised clustering, a critical component of exploratory data analysis, significantly underdeveloped by existing Auto-ML frameworks. Current approaches often limit their scope to dataset characteristics, neglecting the crucial influence of algorithmic suitability (e.g., robustness to outliers) and user-defined requirements (e.g., interpretability needs). This oversight leads to suboptimal clustering outcomes, particularly when dealing with complex, high-dimensional, or noisy data. To address these limitations, this research introduces SOL-Auto-Clust, a novel end-to-end automated clustering framework that makes a key contribution by holistically integrating three fundamental dimensions: inherent data characteristics, intrinsic algorithmic traits, and explicit user-defined objectives. By employing a meta-feature architecture, SOL-Auto-Clust dynamically generates customized clustering pipelines, addressing both data intricacies and real-world application requirements. Extensive evaluation across diverse datasets highlights the framework's ability to simplify clustering processes and produce reliable, insightful outcomes, marking a significant step towards human-aligned Auto-ML for unsupervised learning.
Keywords:	Automated Machine Learning (Auto-ML), Automated Clustering, Unsupervised Learning, CASH
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	AN OPTIMIZED ENSEMBLE MODEL FOR CARDIOVASCULAR WITH DIABETES DISEASE PREDICTION USING CGAN-AUGMENTED DATA
Author:	MUNI BALAJI THUMU , N. BALAJIRAJA , MUHAMMED YOUSOOF
Abstract:	Cardiovascular disease (CVD) holds the position as the main killer worldwide in diabetic populations thus underlining the importance of accurate predictive tools. The inability of traditional statistical methods to adapt to data limitations alongside poor handling of clinical data imbalance leads to unsuccessful risk assessment. Deep learning solutions demonstrate promising results, yet they confront expensive computations and insufficient feature background understanding in addition to lacking interpretability features. The research introduces DFE-CVRP as a cardiovascular risk prediction system which merges expert models tailored for specific features and implements dynamic ensemble control with adaptive data balancing techniques. The performance evaluation determines if a lightweight ensemble model optimized dynamically improves CVD risk prediction results when processing structured clinical data. The method combines EfficientNet architectures which were optimized using Successive Halving and Population-Based Training methods and Conditional Generative Adversarial Networks to balance and improve feature diversity for the dataset. The performance of DFE-CVRP exceeds conventional machine learning techniques together with baseline deep learning architectures such as CCGLSTM when used on structured health databases. The algorithm reaches 98.2% accuracy and 97.8% precision combined with 98.4% recall while obtaining 98.1% F1-score and 98.6% AUC-ROC. The effectiveness of dynamic ensemble learning and data augmentation strategies for improved cardiovascular healthcare diagnosis has been confirmed through the study findings. The proposed predictive framework offers interpretability and scalability as well as affordable resource utilization that creates substantial value for future clinical decision systems leveraging patient-specific data.
Keywords:	Diabetes, Cardiovascular Disease Prediction, GAN, Deep Learning, Machine Learning.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	DEEP LEARNING APPROACHES FOR THE DEVELOPMENT OF INSECT- AND MOULD-RESISTANT PAINTS: AI-DRIVEN FORMULATION AND OPTIMIZATION
Author:	S. HEMALATHA , DR.P.ARIVUBRAKAN ,PONNURU ANUSHA , SURYA LAKSHMI KANTHAM VINTI ,JYOTI D. SHENDAGE , PRAMODKUMAR H KULKARNI
Abstract:	Paint coatings serve as the first line of defense against environmental degradation, yet microbial infestation and insect adhesion continue to pose significant challenges, leading to structural damage and health-related risks. Traditional antifungal and insect-repellent solutions often depend on chemical additives that raise environmental and health concerns. This study investigates the potential of deep learning to optimize paint formulations for enhanced resistance to mould and insect infestation. We introduce a novel AI-driven framework that integrates convolutional neural networks (CNNs) to detect microbial growth patterns, recurrent neural networks (RNNs) to model temporal environmental influences, and generative adversarial networks (GANs) to simulate and generate optimized paint formulations. The system is trained on a comprehensive dataset comprising spectral and microscopic imagery, chemical composition data, and environmental conditions. Our results indicate that the proposed deep learning models outperform conventional heuristic-based methods in identifying effective resistance-enhancing formulations. These findings underscore the transformative role of artificial intelligence in advancing material science and promote the development of eco-friendly, self-adaptive coatings. Future work will involve real-world testing and the integration of IoT-enabled sensors for dynamic resistance management.
Keywords:	Deep Learning, Insect-Resistant Coatings,Mould-Resistant Paints, Convolutional Neural Networks (CNNs),Generative Adversarial Networks (GANs),Antimicrobial Coatings, Smart Paints, IoT-Integrated Coatings, Reinforcement Learning, Material Science AI
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	THE CONSCIOUSNESS SIMULATION GAP: EVALUATING AND BENCHMARKING AI MODELS THROUGH FUNCTIONAL DECOMPOSITION
Author:	MYKHAILO ZHYLIN, TAMARA HOVORUN, BILAL ALIZADE, MAKSYM KOVALENKO, ALLA LYTVYNCHUK
Abstract:	The relevance of the research is determined by the need to study the consciousness of neural networks and the possibility of developing artificial self-awareness. The aim of the article is to investigate the main functional elements of models of consciousness in artificial intelligence (AI). The study employed such methods as the Turing test, Context-driven Testing and analysis of generation models. F1-score, Accuracy, t-test were used for statistical analysis. The reliability of the selected methods was checked by Test-Retest Reliability. The results were obtained that demonstrate the key aspects of the functioning of artificial consciousness models. GPT-4 shows the highest accuracy (92%) and F1-score (0.91), but has difficulties with complex logic problems. AlphaZero has the lowest accuracy (85%) and has trouble understanding abstract concepts. IBM Watson shows medium performance, but does not recognize irony well. DeepMind’s Gato is 90% accurate and wrong on coreference problems. The resulting analysis showed that modern models, such as GPT-4, have a high level of development of perception and attention, which contributes to the effective processing of natural language. However, the question of true consciousness and self-awareness of AI remains open, requiring further research. Understanding the functional components of consciousness is important for the development of ethical norms in the field of AI. Therefore, it is necessary to improve the algorithms to grade up the cognitive functions of the models. Prospects for future research in neural network consciousness include an in-depth study of the mechanisms that provide true consciousness and self-awareness in artificial systems.
Keywords:	Artificial Consciousness, Neural Networks, Dialogue Model, GPT-4, Neuroscience.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	USING VARIANTS OF GENETIC ALGORITHM AND LEARNABLE EVOLUTION MODEL TO SOLVE RESOURCE-CONSTRAINED PROJECT SCHEDULING PROBLEM
Author:	GAMAL ALSHORBAGY, MOHAMED EL-DOSUKY
Abstract:	The resource-constrained project scheduling problem (RCPSP) is a complex scheduling challenge as it is proven to be NP-hard. The Learnable Evolution Model (LEM) is a non-Darwinian evolutionary approach that speeds up convergence using machine learning instead of crossover. It classifies individuals into high-performance (H-group) and low-performance (L-group) based on fitness, learns distinguishing features, and generates new individuals through an instantiation step. To ensure diversity, LEM applies mutation as a Darwinian component, making it more efficient than traditional evolutionary methods. This paper proposes a new approach, which attempts variants of genetic algorithms and LEM, aiming to tackle issues in generating Gantt charts for big cases.
Keywords:	Scheduling, RCPSP, Learnable evolution model, Genetic Algorithm
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	FROM STATIC DISPLAYS TO INTERACTIVE AR: EVALUATING THE EFFECTIVENESS OF AN AR APP FOR GEN Z AND ALPHA IN MUSEUM PUSAKA
Author:	REHMAN ULLAH KHAN, YIN BEE OON, AMALIA BT MADIHIE, MOHAMAD HARDYMAN BIN BARAWI, IDA JULIANA HUTASUHUT, HARI NUGRAHA RANUDINATA, DESI DWI KRISTANTO
Abstract:	Museums play a crucial role in preserving and disseminating cultural heritage; however, they often struggle to engage modern audiences who seek immersive and interactive experiences. Traditional static displays fail to engage visitors or transfer cultural values to the new generations, specifically the Gens Z and Alpha. This research explores the implementation of Augmented Reality (AR) in museum settings to enhance visitor engagement and learning effectiveness, with a particular focus on the Museum Pusaka at Taman Mini Indonesia Indah in Jakarta. An applied empirical research design was utilized, two different studies involving a total of 40 and 86 students who were divided into a control or an experimental group. The experimental group used an AR mobile application that contains 3D models and interactive content, while the control group used traditional printed materials. To measure the learning outcomes and engagement, pre-and post-tests were conducted, and the data were analysed and compared using paired sample t-tests and ANCOVA. The results showed significant improvements in both learning and engagement among participants using the AR mobile application. These findings indicate that AR has the potential to transform the museum exhibit more engaging, interactive, and impactful.
Keywords:	Augmented Reality, Museum Studies, Educational Technology, User Experience, Learning Outcomes
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	MULTI HEAD ATTENTION-BASED LSTM AND GRADIENT-WEIGHTED CLASS ACTIVATION MAPPING FOR BRAIN TUMOR DETECTION USING MRI IMAGES
Author:	PRATHIMA DEVADAS, G. MATHIVANAN
Abstract:	The presented of attention-based architectures in medical imaging has ushered in a novel era of precision diagnostics, mainly for the identification and classification of brain tumors. This research developing a novel knowledge distillation method that employs a tripartite attention mechanism within transformer encoder models to identify various brain tumor types utilizing magnetic resonance imaging (MRI). This study offerings a unique method for brain tumor identification, integrating Multi-Head Attention-based Long Short-Term Memory (MHA-LSTM) networks with Gradient-weighted Class Activation Mapping (Grad-CAM). The MHA-LSTM design utilizes multi-head attention to capture complex spatial-temporal relationships across consecutive MRI slices, enabling the model to focus on the most critical features. Grad-CAM is incorporated to provide visual explanations by highlighting key regions contributing to the model's predictions, improving both interpretability and clinical relevance. Experimental results demonstrate that the suggested technique surpasses conventional LSTM models in terms of accuracy, sensitivity, and specificity. Moreover, the Grad-CAM visualizations offer valuable insights into the model's decision-making process, fostering better understanding and facilitating future clinical validation. This method presented a robust and interpretable solution for brain tumor identification, advancing the application of deep learning in medical imaging.
Keywords:	Brain Tumor Detection, Deep Learning, Grad-CAM, Interpretability, Long Short-Term Memory, Magnetic resonance imaging, Medical Imaging, Multi-Head Attention, Neural Networks, Spatiotemporal Modeling.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	NEXT-GEN SECURITY: LEVERAGING DNA CRYPTOGRAPHY FOR ROBUST ENCRYPTION
Author:	GURU PRAKASH B , SIVA T , SHUNMUGASUNDARAM S , MARIAPPAN E , ANNA LAKSHMI A , RAMNATH MUTHUSAMY
Abstract:	Cloud computing is the popular growing technology that provides services through the internet for data sharing and storage, access. Cryptography is the study of protecting the information by using algorithms, codes so that the intended users can view the data. Cryptography plays a vital role while transmitting data through networks and it’s very important to ensure the confidentiality of the data. In this paper, to achieve and enhance the confidentiality of the data, DNA cryptography has been proposed. DNA cryptography is used to enhance the security of the data which is purely based on the nucleotide of DNA. The proposed modernized DNA cryptography algorithm is implemented using the .net framework and examples are also given with screenshots for the conclusion.
Keywords:	Cloud Computing, Secure Communication, DNA Cryptography, Amino Acid Tables, Dynamic Key Generation.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	ADVANCED CNN-BASED FRAMEWORKS FOR ROBUST AND EXPLAINABLE BREAST CANCER DIAGNOSIS ACROSS MULTI-MODAL IMAGING DATASETS
Author:	Dr. ALURI BRAHMAREDDY , Dr. MERCY PAUL SELVAN
Abstract:	Breast cancer stands as the principal cause triggering female fatalities worldwide hence requiring immediate diagnostic solutions which produce precise and timely results. Expanding from its demonstrated potential deep learning technologies function with constraints stemming from single-type medical image analysis as well as inadequate transparency regarding decision-making operations. This research examines deficient diagnostic methods because they fail to effectively link multimodal medical images with healthcare parameters to diagnose breast cancer both precisely and explainable to medical staff. The predictive model uses a combination of mammograms, ultrasounds and MRIs together with histopathological images and structured clinical data features like age, breast density, lesion size, genetic marker scores and tumor stage to achieve better predictive results and stronger model generalization. This study eliminates the present knowledge gaps through a novel Multi-Modal Explainable Convolutional Neural Network (MME-CNN) framework that unites mammograms with MRIs and ultrasounds, histopathological images and structured clinical data containing age, lesion size, breast density, genetic markers and tumor stage information. Grad-CAM visualizations served within the model as an interpretability tool that shows doctors how the predictions were made. The experimental analysis shows the model achieved a perfect validation accuracy of 100% while needing only four epochs to complete training. It reduced training loss from 0.6975 to 0.2551 and established a validation loss at 0.0886. Real-world clinical implementations benefit from this framework because it shows good universal applicability and quick calculation rates and better explanation capabilities. The future development will concentrate on conducting extensive validity tests alongside EHR system combinations to enable widespread precision oncology implementation.
Keywords:	Breast Cancer Diagnosis, Convolutional Neural Networks, Multi-Modal Imaging, Robust Classification, Explainability, Precision Medicine, Medical Imaging Analysis And Grad-CAM Visualization .
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	LEVERAGING DEEP LEARNING FOR REAL-TIME, CONTINUOUS MONITORING AND PREDICTION OF SEPSIS IN ICU PATIENTS USING MULTISENSORY DATA
Author:	BHOOMPALLY VENKATESH, DR. PRADEEP K R, DR. Y.RAMADEVI, DR VISHWA KIRAN S, KAPARTHI UDAY
Abstract:	Sepsis is a life-threatening condition caused by the body's extreme response to infection, which may end fatally, with high death rates, particularly in the critical care unit, associated with multi-organ failure. Physiological data in ICU patients is highly non-stationary and complex, significantly challenging early detection. Standard machine learning models and traditional scoring systems fail to learn spatial and temporal patterns and accurately provide poor early detection performance. To overcome these limitations, we present a new framework called SepsisNet, a deep-learning model for real-time continuous sepsis prediction from multisensory ICU data. Our proposed attention-based CNN BiLSTM model incorporates CNN for spatial feature extraction and BiLSTM networks for temporal sequence modeling and adds an attention mechanism to emphasize the most informative physiological features for classification. On the benchmark dataset MIMIC-III, experimental results show that SepsisNet achieves 98.68% accuracy, surpassing the baseline models, including Logistic Regression, Random Forest, SVM, LSTM, and standard CNN. The ablation study also reinforces the importance of each architectural component. We demonstrate that SepsisNet has the potential to serve as a reliable, interpretable, and computationally efficient sepsis predictor, thereby enabling real-time clinical decision support in ICU settings. This research will help enhance sepsis detection as well as play a significant role in early medical treatment, which is significantly required to reduce fatalities caused by sepsis.
Keywords:	Sepsis Prediction, Deep Learning, CNN, BiLSTM, Attention Mechanism
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	PREDICTIVE MODELING AND MULTIVARIATE ANALYSIS OF CORE FOOD SECURITY INDICATORS IN MOROCCO
Author:	MEHDI RAHMAOUI , ACHRAF CHAKIR BARAKA, AHSSAINE BOURAKADI , NADA YAMOUL , HAMID KHALIFI , ABDELLATIF BOUR
Abstract:	This study presents a multivariate analysis of key food security indicators in Morocco between 2000 and 2023. The first section introduces the theoretical framework of linear regression and principal component analysis (PCA). A linear regression model was then applied to examine the relationship between the prevalence of undernourishment and the Consumer Price Index (CPI in %), yielding a high coefficient of determination (R² = 95%). The regression model demonstrated that a one-unit increase in CPI leads to a 0.04% rise in undernourishment prevalence. The PCA of food security indicators highlights two distinct dimensions that shape nutritional outcomes. The first principal component, accounting for most of the variance, shows strong positive correlations with dietary energy adequacy (0.92), minimum caloric requirements (0.98), and per capita GDP (0.96), while inversely relating to food supply instability (-0.81). This axis essentially measures a nation's economic strength and food system resilience. The second component exclusively tracks undernourishment metrics, with near-perfect alignment to both the rate (0.98) and absolute numbers (1.00) of underfed populations. This dimension directly reflects the human toll of food insecurity. Importantly, these components are statistically independent (orthogonal), which reveals an important reality of policies: economic growth and enhancements to the food system (Dimension 1) do not translate into reduced hunger (Dimension 2). This separation reinforces the necessity to engage in bi-focal approaches to fighting food insecurity - macroeconomic policies strengthen the food system while humanitarian action focuses on nutrition for specific populations. These results show that food security exists on different levels and requires solutions that target the systemic economic level and humanitarian levels to address the multi-faceted concept of hunger and malnutrition. Ultimately, ARIMA forecasting is utilized to forecast trends for Dietary Energy Adequacy, Prevalence of Undernourishment (% of population), and Number of Undernourished People between 2025 and 2028. Forecasts suggests a gradual improvement in the following areas: Dietary Energy Adequacy rising from143.1 to 145.9 kcal/cap/day, Prevalence of Undernourishment drops from 5.88% to 5.19%, and Number of Undernourished People falling from 2.09 to 1.96 million. This evidence offers substantial information for policy makers to improve food security policies in Morocco.
Keywords:	Food Security, Morocco, Linear Regression, Principal Component Analysis (PCA), ARIMA Models, Forecasting, Undernourishment.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	BLOCKCHAIN TECHNOLOGY AND ITS IMPACT ON FINANCIAL REPORTING IN THE DIGITAL ACCOUNTING ERA
Author:	AHMAD ALNAIMAT MOHAMMAD , OLEKSANDR CHUMAK , MYKYTA ARTEMCHUK , ALONA KHMELIUK , SVITLANA SKRYPNYK
Abstract:	The study is relevant, as blockchain technologies transform financial reporting, increasing its transparency, reliability, and data processing speed, reducing costs and minimizing risks, which requires further analysis. However, despite numerous studies on blockchain applications, a knowledge gap exists in understanding its comprehensive impact on financial reporting processes and the development of integration models with other digital technologies. The aim of the study is to determine the impact of blockchain technologies on the processes of preparing and submitting financial statements, as well as determining the prospects for using this technology in accounting in the digital age. The research employed the following methods: content analysis of modern blockchain systems, comparative analysis of financial indicators of companies that use blockchain, as well as economic and statistical modelling. The impact of blockchain technologies was assessed through quantitative analysis, including descriptive statistics, analysis of variance (ANOVA), correlation analysis (Pearson and Spearman coefficients), regression analysis, cluster analysis and hypothesis testing (t-test, Mann-Whitney U-test). The calculations were performed using SPSS, Stata, and Python software (Pandas, Statsmodels, Scikit-learn). The results confirm that the implementation of blockchain technologies increases the efficiency of financial reporting, reducing operating costs by 15–20% and reducing audit costs by 25–30%. Smart contracts minimize errors by 18%, and the average processing time for financial transactions decreased from 48 to 5 hours. In the financial sector, costs were reduced by 30%, and transaction processing time by 85%. The academic novelty of the study lies in the comprehensive analysis of the application of blockchain technologies to increase the transparency and reliability of financial reporting in a global context, as well as the creation of new knowledge through statistical analysis and practical assessment of blockchain's effectiveness. The prospects for further research include the development of models for integrating blockchain with other digital technologies, such as artificial intelligence (AI) and Big Data, as well as assessing the long-term economic consequences of using blockchain in the financial sector.
Keywords:	Blockchain, Financial Reporting, Digital Accounting, Smart Contracts, Transparency, International Standards, Automation.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	ADVANCED RIDGE REGRESSION USING IMFT MODEL FOR RSU DESIGN IN VEHICULAR NETWORKS
Author:	KRISHNA KOMARAM , NAGARJUNA KARYEMSETTY
Abstract:	In the realm of Vehicular Ad Hoc Networks (VANETs), Roadside Units (RSUs) play a pivotal role in enhancing communication, data processing, and predictive analytics. This paper introduces a novel hybrid design that integrates Ridge Regression and XG-Boost algorithms to optimize the data processing and prediction capabilities of RSUs, aimed at improving traffic management and safety applications. The hybrid framework with IMFT (inter-intra mod filter) algorithm employs Ridge Regression for robust initial data processing, minimizing overfitting and ensuring reliability in the noisy, dynamic environment of vehicular data. Feature extraction with IMFT is utilized to encompass the relevant features before utilizing Ridge-Model for high-accuracy predictions, leveraging its gradient boosting capabilities to facilitate timely interventions and optimize traffic flow. Furthermore, the architecture of the RSU is expanded to include essential units such as communication modules, data storage, and user interface components, all functioning cohesively to create a comprehensive system. With the proposed IMFT design we have incorporated extensive simulations with K-fold loss to demonstrate that the proposed IMFT with Ridge Model design significantly enhances prediction accuracy and processing efficiency compared to traditional methods (Elastic Net.) with more than 98% of improved R2-score. By optimizing the operational capabilities of RSUs in VANETs, this work contributes to the development of smarter and safer urban mobility solutions, paving the way for more effective traffic management and improved vehicular safety.
Keywords:	Vehicular Ad Hoc Networks (VANETs), Roadside Units (RSUs), Security Protocols, Energy Management, ML (Machine Learning), Ridge Regression.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	SELF-RELIANT RESIDUAL NETWORK BASED DEEP LEARNING FRAMEWORK FOR MELANOMA SKIN DISEASE DETECTION
Author:	V. RADHIKA , A. MUTHUCHUDAR , M. LINGARAJ
Abstract:	Melanoma is one of the deadliest types of skin cancer and one of the most aggressive that may be if caught late. While traditional approaches may have their limits, an accurate diagnosis is vital for patient survival. Thanks to its capacity to understand intricate patterns from massive datasets, deep learning has evolved as a potential method for automated melanoma diagnosis. The challenges persist, including overfitting, instability during training, and difficulties in handling nonlinearities, which can hinder accurate predictions. To address these challenges, Self-Reliant ResNet (SR-ResNet) has been proposed. This enhanced version of ResNet integrates Zoutendijk’s Method, a nonlinear optimization technique, to optimize weight updates and improve convergence. SR-ResNet features a series of residual blocks where Zoutendijk’s Method refines the learning process, ensuring stability and efficient training, even in deeper networks. The network’s architecture has been designed to enhance performance and generalization. The proposed SR-ResNet has been evaluated using a dataset of 10,000 Melanoma Skin Cancer images. The results demonstrate significant improvements in classification accuracy, achieving a high precision rate with reduced overfitting. SR-ResNet outperforms traditional models, establishing itself as a robust tool for melanoma diagnosis.
Keywords:	Melanoma Skin Cancer, Deep Learning, SR-ResNet, Zoutendijk’s Method, Classification Accuracy
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	DIFT-VAR: A DYNAMIC MULTI-LAYER FRAMEWORK FOR DEVICE FINGERPRINTING OF IDENTICAL DEVICES
Author:	MANOJ KUMAR VEMULA , KAILA SHAHU CHATRAPATI
Abstract:	Device fingerprinting is a powerful technique for identifying devices in an IoT environment, offering multiple advantages such as enhanced security through device authentication, improved network management by monitoring device behaviors, and anomaly detection for identifying unauthorized or compromised devices. The majority of recent fingerprinting schemes consider a heterogeneous device environment and use different machine learning techniques to identify devices using network traffic, signal-level information, radio frequency characteristics, etc. However, fingerprinting devices of the same make and model is a significant challenge in modern IoT environments, where many devices often share identical hardware and software configurations. Existing techniques cannot reliably differentiate identical devices as they lack sufficient data. This paper proposes a novel approach for Device Identification and Fingerprinting with Time-Variant Adaptive Recognition (DIFT-VAR) based on multi-layer, time-varying feature extraction. We construct dynamic fingerprints that uniquely identify each device by monitoring and fusing features such as probe request behavior, clock skew, transport layer characteristics, and radio signal metrics over time. We utilize machine learning algorithms such as Random Forests to classify devices based on these dynamic fingerprints. We further propose the use of dynamic time warping (DTW) for feature alignment and classification. Experimental results demonstrate the efficacy of our approach in distinguishing identical devices with an accuracy of over 97% using standard machine learning metrics.
Keywords:	Device fingerprinting, IoT Security, Dynamic time warping (DTW), Time-variant feature extraction, Machine learning for IoT security.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	SCALABILITY AND EFFICIENCY OF CLUSTERING ALGORITHMS FOR LARGE-SCALE IoT DATA: A COMPARATIVE ANALYSIS
Author:	PRABHAT DAS, KARTHIK KOVURI, SAJAL SAHA
Abstract:	This research investigates the scalability and efficiency of clustering algorithms applied to large-scale Internet of Things (IoT) data. A comprehensive evaluation is conducted on fourteen clustering algorithms—Affinity Propagation, Agglomerative, BIRCH, Bisecting K-Means, DBSCAN, Fuzzy C-Means, Gaussian Mixtures, HDBSCAN, K-Means, Mean-Shift, OPTICS, Overlapping K-Means, Spectral Clustering, and Ward-Hierarchical—across datasets ranging from 40,000 to 100,000 sensor readings. The study systematically analyzes execution time and clustering performance to determine their suitability for large-scale IoT applications. Results indicate that K-Means, Ward-Hierarchical, and BIRCH exhibit strong scalability and computational efficiency, whereas Affinity Propagation and Spectral Clustering face significant challenges with increasing dataset size. These findings provide valuable guidance for selecting optimal clustering techniques in IoT-based data analytics, considering factors such as computational constraints, dataset characteristics, and clustering granularity.
Keywords:	Clustering Algorithms, IoT Data Clustering, Comparative Analysis, Sensor Data Analysis, Bibliometric Analysis, Machine Learning in IoT, Multi-Dimensional Data Clustering
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	HOW CAN EXPERT CONSENSUS METHODS ENHANCE THE DESIGN OF IMMERSIVE LEARNING PRACTICAL MODELS FOR DEAF OR HARD-OF-HEARING STUDENTS IN TVET?
Author:	RINI HAFZAH BINTI ABDUL RAHIM, DINNA NINA BINTI MOHD NIZAM, NUR FARAHA BINTI MOHD NAIM, ASLINA BAHARUM
Abstract:	Traditional auditory-based teaching approaches limit the effectiveness of practical skills acquisition for Deaf or Hard-of-Hearing (DHH) students in Technical and Vocational Education and Training (TVET). Despite increased interest in immersive technologies like augmented reality (AR), the field lacks validated, inclusive instructional models tailored to DHH learners. This study addresses this gap by integrating the Nominal Group Technique (NGT) and Fuzzy Delphi Method (FDM) to design and validate an Immersive Learning Practical Skills (ILPS) model. The novelty lies in the combined use of NGT and FDM for consensus-building among experts in AR, gamification, and DHH education—an approach not commonly applied in inclusive model development. Results revealed a strong expert consensus (>97%) on 15 items across three core constructs: Learning Input Medium, Practical Skills Module, and AR Gamification Features. This research offers a replicable and participatory model development process and introduces a validated framework for inclusive immersive learning in TVET. The study contributes new knowledge by demonstrating how expert-driven methods can operationalize inclusive pedagogy through immersive technologies. This study demonstrates how combining FDM and NGT may successfully evaluate inclusive design elements for immersive learning. The results support the development of a practical skills model with a DHH focus and provide a repeatable framework for inclusive curriculum co-creation. This combination strengthen consensus among 11 panel of experts and according to the study's findings, the NGT and FDM approach has made it simple and quick for researchers to confirm crucial details that should be highlighted. To help DHH students learn more effectively, it is advised that more research be done in collaboration with course designers. To provide a scalable approach for developing immersive, accessible learning environments in specialized educational contexts, this study hopes to demonstrate how effectively NGT and FDM collaborate for inclusive instructional design.
Keywords:	Educational Technology, Teaching And Learning, Hearing Impaired, Higher Education, Model Development.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	CAN DIGITAL TRANSPARENCY TOOLS SYSTEMATICALLY REDUCE CORRUPTION IN GOVERNMENT? EVIDENCE FROM ESTONIA, UKRAINE AND BRAZIL
Author:	OLEKSANDR KOTUKOV, DMYTRO KARAMYSHEV, TETIANA KOTUKOVA, ALINA CHERNOIVANENKO, ARTEM SERENOK
Abstract:	This study addresses a critical gap in the existing literature, which has primarily focused on general transparency rather than the specific impact of digital tools in various political and institutional contexts. Despite the proliferation of e-governance initiatives, there is limited empirical research systematically comparing the effectiveness of digital transparency in mitigating corruption across multiple countries.To bridge this gap, we investigate how digital governance platforms influence institutional accountability and reduce corruption in Ukraine, Estonia, and Brazil. Using a combination of econometric modeling, comparative case analysis, and time-series analysis across 120 government institutions, we demonstrate that digital transparency tools—particularly open data platforms, e-procurement systems, and AI-driven fraud detection—are associated with statistically significant reductions in corruption rates. Estonia, with its mature digital ecosystem, achieved a 39% reduction in corruption, followed by Ukraine (28%) and Brazil (16%). The novelty of this study lies in its comparative design, the integration of AI analytics, and the identification of conditions under which digital transparency tools are most effective. Our findings provide actionable insights for policymakers, emphasizing the need for robust digital infrastructure, legal mandates for data openness, and civic engagement to maximize anti-corruption outcomes. This research contributes new empirical knowledge on how digital transparency tools can transform public administration and strengthen institutional integrity.
Keywords:	Digital Transparency; Corruption Reduction; E-Governance; Institutional Accountability; Public Administration
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	SQUIRREL SEARCH GRADIENT OPTIMIZED DEEP BELIEF NETWORK CLASSIFIER FOR THYROID DISEASE PREDICTION
Author:	R.VANITHA , Dr.K. PERUMAL
Abstract:	Thyroid disease is a range of disorders that affect the thyroid gland, a butterfly-shaped organ located in the neck responsible for producing hormones that regulate metabolism, energy levels, and overall bodily functions. Early detection and management of thyroid disease are crucial, as untreated conditions leads to severe complications, including cardiovascular issues, infertility, and metabolic disorders. Advanced diagnostic methods, including machine learning and deep learning techniques, are increasingly used to improve the accuracy and timeliness of thyroid disease detection, facilitating better treatment outcomes. But, severity of thyroid disease prediction accuracy with minimal time is major challenging issues. In order to improve the accuracy of thyroid disease prediction, a novel Squirrel Search Gradient Optimized Deep Belief Neural Classifier (SSGODBNC) model is developed with minimal time consumption. The proposed Deep Belief Network (DBN) is a fully connected artificial feed-forward deep learning method comprising two visible layers such as the input and output layer and multiple hidden layers for processing the given input. In the layer-by-layer process, the first hidden layer receives weighted input and performs data preprocessing. Then extracting significant features and eliminates the insignificant features from the dataset using the Sparse Autoencoder model. These selected significant features are utilized to classify the severity level of thyroid disease using Sokal–Michener’s simple matching method. During fine-tuning, error back-propagation algorithms adjust the hyperparameters using Squirrel Search Gradient Optimization to increase the accuracy of thyroid disease classification. This optimized fine-tuning process significantly enhances the performance of the deep belief network and improves overall learning efficiency in classification tasks. Finally, the accurate thyroid disease severity prediction results with minimal error are obtained at the output layer. Experimental assessment is conducted with different evaluation metrics such as Accuracy, Precision, Recall, F1-score, specificity and Thyroid disease prediction time. The observed result shows the effectiveness of the proposed SSGODBNC model with higher accuracy in thyroid disease prediction with minimum time than the existing methods.
Keywords:	Thyroid Disease Prediction, Deep Belief Network, Fine-Tuning, Adaptive Gradient Method, Squirrel Search Gradient Optimization, Sokal–Michener’s Simple Matching Method.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	A COMPUTATIONAL MODEL FOR TEA LEAF PRICE PREDICTION BASED ON QUALITY FACTORS USING HYBRID MACHINE LEARNING TECHNIQUES
Author:	IRA GABA , B RAMAMURTHY
Abstract:	This document reflects the effort made to calculate and identify the grade of the tea leaves based on the assessment of the leaves' size and color. The leaves were classified based on their severity with the help of HSV. The leaves were further classified using the k prototypes clustering once their length and width were established. The leaves were then further categorized in line with that. Light, medium, and dark are the three-color categories into which it belongs. The leaves were further sorted according to their quality so that the farmer could sell the produce at a better price. With the machine learning method for the categorization part, we were able to show its values. All of the healthy leaves were considered in a different dataset, and the images were obtained using the feature selection method. The length and width of each individual leaf, along with its color and shape, were then measured using those leaves. We were able to differentiate between the various leaf grades based on the findings. The healthy leaves were separated from the diseased leaves using the textual features. Additionally, we were able to use the other criteria to obtain higher-grade leaves.
Keywords:	Image Pre-Processing, Feature Selection, Classification, HSV, Color Parameters, K-Prototypes Clustering.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	LEVERAGING SPEECH FOR DYNAMIC IMAGE CAPTIONING: A MOBILENETV3 AND LSTM APPROACH
Author:	PREETY SINGH, NAGA DURGA SAILE K, TAKKEDU MALATHI, T RAVI, 5DIPAK J DAHIGAONKAR, CHUNDURI LAVANYA
Abstract:	This paper presents a novel way of generating captions for images using an automatic image captioning system; the proposed model combines MobileNetV3 and LSTM to create captions that are accurate and relevant to the image being depicted. MobileNetV3 acts as a feature extractor as it extracts vital components of pictures at a reasonable computational cost. These features are then taken to the LSTM network and from it, descriptive captions from visual context are made. As a measure that improves user access, the generated captions are further translated to sound using Google Text-to-Speech (GTTS), which is especially important for the visually impaired and other hand-free users. Cross-sectional experimental assessments of the performances of the proposed model on the Flickr8k dataset further validate the impressive usefulness of the proposed model for generating faithful and comprehensive descriptions for media items that are potentially useful in assistive technologies, media organizing, and interactive systems.
Keywords:	MobileNetV3, Long Short-Term Memory (LSTM), Google Text-to-Speech (GTTS)
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	EDGE LEVEL COMPLEX EVENT PROCESSING AND OPTIMISATION BASED FOG LEVEL LOAD BALANCING MECHANISM IN SOFTWARE DEFINED NETWORK
Author:	SARKARSINHA HARSINHA RAJPUT, DR. MANOJ EKNATH PATIL
Abstract:	This study uses an efficient transmission model based on the Hybrid Meta-heuristic Model to enhance data transfer by reducing time complexity. Initially, data is moved into the Complex Event Processing (CEP), which is positioned between the fog layer and the IoT layer. In edge IoT devices, complex event processing comprises real-time analysis, correlation, and interpretation of continuous data streams generated by sensors and edge devices. It seeks to identify significant trends or intricate occurrences in various data streams in order to facilitate quick decisions or immediate reactions. After CEP, a multi-tier priority queue-based model is used to attain priority-aware task scheduling. After the arrival of all the tasks, each task is sorted into slots based on its category. High-priority tasks are completed first due to their preference over lower-priority slots. A software-defined network's optimal resource utilization and task response time are guaranteed by an effective load-balancing method called Hybrid Pigeon Cat Search Optimization Algorithms (HPC_SOA). Arranging tasks based on their availability, capacity, proximity, and energy efficiency may optimize the fog nodes' resource utilization and energy usage. In the evaluation, the proposed approach has consumed 22051 Kw/h of energy.
Keywords:	Complex Event Processing, Priority Queue Approach, Pigeon Optimization, Cat Search Optimization, Software-Defined Network.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	TO-MULTILONTOLOGY & MPCO: A METHODOLOGY FOR DEVELOPING MULTILINGUAL ONTOLOGIES & A LEGAL ONTOLOGY OF THE PENAL CODE
Author:	ISMAHANE KOURTIN
Abstract:	Ontologies are among the techniques introduced by artificial intelligence in the early 1990s to enable better organization and semantic representation of information. They have the potential to play a crucial role in the design of question-answering systems and content comprehension by organizing and structuring the data they present. Multilingual ontologies are both language-independent and capable of supporting multiple languages, offering significant potential for querying and understanding knowledge in multicultural and multilingual environments. Although several ontology development methodologies exist, they provide the necessary elements for ontology construction without clearly demonstrating how to implement them or specifying the models to guide the development process, particularly for multilingual ontologies. Indeed, existing methodologies summarize the development of ontologies as a mere enumeration of important terms, followed by the definition of classes and their hierarchy, the definition of properties and their facets, and finally the creation of instances—without showing users the approach or method that could guide them in choosing terms, defining classes, the hierarchy, and properties, or in demonstrating how to build multilingual ontologies. In addition, there is a lack of models that allow for representing ontology data in a way that guides its development and documentation. This article proposes a customized methodology, TO-MULTILONTOLOGY, which covers aspects from the specification phase to the validation and evaluation phase, offering a detailed implementation process with clearly defined steps to guide and simplify the task of building multilingual ontologies. The proposed methodology also addresses one of the main obstacles to effective knowledge sharing: the inadequate documentation of existing ontologies. It provides powerful tools and models that not only document the ontology but also guide its development. This methodology will be explained and applied in the development of a multilingual legal ontology, MPCO (Multilingual Penal Code Ontology), in French and Arabic, for the Moroccan government's Penal Code. The constructed ontology can play a significant role in information retrieval and in learning about the penal code. It can also serve as a reference for the development of similar penal law ontologies.
Keywords:	Legal Ontologies, Ontology Development, Ontology Design, Multilingual Ontologies, Knowledge Representation And Modeling, Ontology Construction And Development Methodologies.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	ENHANCING ROBUSTNESS IN MEDICAL QUESTION ANSWERING SYSTEMS WITH NOVEL DEFENSE MODELS AGAINST ADVERSARIAL ATTACKS
Author:	ATRAB A. ABD EL-AZIZ, REDA A EL-KHORIBI, AND NOUR ELDEEN KHALIFA
Abstract:	Medical Question Answering (MQA) systems play a critical role in supporting accurate medical diagnoses and healthcare decision-making. However, they are increasingly vulnerable to adversarial text attacks. These attacks subtly alter input questions and lead to incorrect outputs. While prior research has extensively explored adversarial defenses for medical images, there remains a significant gap in protection strategies for text-based MQA systems. To the best of our knowledge, this paper is the first to propose and evaluate defense mechanisms specifically designed to secure MQA systems against these attacks. We introduce three novel defense models that address both word-level (synonym substitution, word deletion) and character-level (random character insertion) attacks targeting the BERT model. The Synonym Substitution Embedding (SSE) Defense Framework combines TF-IDF ranking with transformer-based synonym embeddings to resist synonym substitution attacks. CosineDefender leverages cosine similarity to detect and neutralize perturbed inputs, while JaccardDefender applies Jaccard similarity to provide robust protection across multiple attack vectors. To validate our approach, we conduct experiments on two medical datasets (Symptom2Disease and Medical Symptoms Text and Audio Classification) and a natural language dataset (AG’s News) for comparative analysis. Our results show that the SSE model reduces the attack success rate on AG’s News from 8.7% to just 0.4%. On medical datasets, CosineDefender significantly lowers attack success rates to 3.4%, 4.3%, and 12.8%, while JaccardDefender consistently achieves the best performance, reducing all attack success rates to around 3.4% and maintaining high classification accuracy. This work introduces a new line of defense for MQA systems. It establishes a baseline for adversarial robustness in the medical NLP domain. It also contributes the first comprehensive evaluation of targeted defense models in this critical area.
Keywords:	Adversarial Attacks, BERT, Medical Question Answer (MQA), Term Frequency-Inverse Document Frequency (TFIDF).
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	HYBRID INTRUSION DETECTION FRAMEWORK FOR MOBILE EDGE COMPUTING
Author:	SUJAN KUMAR DAS , MOHAMED EL-DOSUKY , SHERIF KAMEL
Abstract:	The growing use of mobile edge computing (MEC) has had a positive impact on user experience and reduced latency. However, this closeness also makes MEC environments vulnerable to a number of security risks. This research article presents an edge-based hybrid intrusion detection system for MEC and the Internet of Things (IoT). The system uses techniques like behavioral analysis, anomaly detection, and signature-based detection, ensuring real-time response and reduced bandwidth usage. The system also addresses challenges in data acquisition and cleaning due to potential threats from malicious users and noise. The model uses smoothing filters, unsupervised learning, and deep learning techniques to detect anomalies and threats, reducing bandwidth. According to the findings, securing MEC environments against changing cyber threats can be accomplished using an edge-based hybrid intrusion detection system.
Keywords:	Mobile edge computing, Intrusion detection, Blockchain, Hybrid intrusion, Machine learning
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	A DYNAMIC IMAGE RETRIEVAL FRAMEWORK BASED ON FUSION-BASED FEATURE EXTRACTION USING DEEP LEARNING
Author:	ALLA TALIB MOHSIN, MOHD SHAFRY BIN MOHD RAHIM
Abstract:	Images are integral to human communication, and with the rapid growth of multimedia data, finding relevant images in large archives has become a significant challenge. Content-Based Image Retrieval (CBIR) offers a solution by retrieving visually similar images based on content rather than textual annotations. Despite its potential, CBIR systems face critical challenges, including irrelevant region detection, sensitivity to variations in brightness and size, and the absence of predefined class information in datasets. To address these issues, this study proposes a CBIR framework that integrates low- and high-level features such as texture, shape, and color for robust image representation. The framework employs wavelet-based Local Ternary Pattern (LTP) for texture extraction and incorporates a dynamic weight allocation mechanism, which adapts to statistical metrics like mean and variance to enhance retrieval accuracy. Comprehensive evaluations of the Corel-1k and Corel-10k datasets demonstrate the method's effectiveness in retrieving visually similar images with high precision. The proposed approach surpasses existing techniques, including CBIR-ANR, OMCBIR, and CNN-QCSO, in terms of precision, memory efficiency, and visual quality. The result of this work for the feature types included the dataset in the Retrieval Performance on the Corel 1K Dataset of fused features, proposing a precision of 88.1, recall of 80.5, and MAP of 88.3, and the Retrieval Performance on the Corel 10K Dataset of fused features proposed the precision of 83.7, recall of 76.2, and MAP of 83.9. This study establishes a promising direction for developing efficient CBIR systems capable of handling large-scale image datasets while improving retrieval performance.
Keywords:	Image Retrieval; CBIR; Features extraction; CNN; LTP.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	EXPLORING THE FACTORS AFFECTING THE DEPLOYMENT OF THE INTERNET OF THINGS IN HEALTHCARE ORGANIZATIONS IN THE UAE USING THE UTAUT MODEL
Author:	NAHIL ABDALLAH
Abstract:	Despite the growing interest in digital healthcare, the adoption of Internet of Things (IoT) technologies in healthcare organizations, remains limited and underexplored, particularly from the patients' perspective. This research investigates the key factors influencing the deployment of IoT-based healthcare devices among end users in public hospitals across the UAE. Drawing from the Unified Theory of Acceptance and Use of Technology (UTAUT), enhanced with constructs identified from existing literature, the study proposes a predictive adoption model. Data was gathered from 231 participants, and structural equation modeling was used to validate both the measurement and structural components of the model. The findings highlight that technological complexity, social influence, perceived health risks, facilitating conditions, perceived security and privacy, and relative advantages significantly shape users' attitudes (ATT), which in turn affect their behavioral intentions (BI) to adopt IoT healthcare devices. The study concludes that addressing these factors is critical for successful IoT implementation in healthcare. It contributes to Information Systems (IS) research by integrating new variables into the UTAUT model and offers practical insights for healthcare decision-makers and technology providers aiming to boost IoT adoption.
Keywords:	Healthcare, Internet of Things, Privacy, Security, UTAUT
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	ROBUST QUANTILE REGRESSION-BASED MACHINE LEARNING FRAMEWORK FOR OUTLIER-RESILIENT TIME SERIES ANALYSIS
Author:	Mr. TUSHAR MEHTA, DR.DHARMENDRA PATEL
Abstract:	Time series analysis is a powerful tool in countless regions, from finance to healthcare, but is often challenged by the presence of outliers that can distort predictions and model output. This article presents a robust quantile regression-based framework for machine learning, which increases the resilience of time series analysis compared to outliers. By using quantile regression, the proposed framework captures the conditional distribution of time series data and provides a more comprehensive understanding of its underlying structure. Machine learning integration improves the ability of models to adapt to complex nonlinear patterns while simultaneously maintaining robustness to anomaly data points. Through extensive research into synthesis and practical datasets, we show that our framework outweighs traditional predictability and trigger resilience methods. The results highlight the possibility of combining quantile regression with machine learning for robust time series analysis, providing promising directions for future research and applications in the environment. Time series forecasts play a key role, especially in financial markets where accurate forecasts are useful for investors and stakeholders. However, traditional models have difficulty recording nonlinear dependencies and are not able to effectively handle outliers. This article presents a robust quantile regression-based machine learning framework for diffusion-preserving time series analysis. It integrates long-term time memory (LSTM), LightGBM, and stacked ensemble models. The proposed ensemble approach uses quantile regression for robust outsourcing processing while combining deep learning strengths to increase base techniques to improve predictive performance. Experimental evaluation of Goldman Sachs BDC, Inc.(GSBD) shared course data demonstrates the advantages of the ensemble model compared to the individual model. The results show that the stacked ensemble model reaches the lowest flipper loss (0.0656), MAE (0.1313), RMSE (0.2185), and the highest R² (0.9778), exceeding LSTM and LightGBM. The results highlight the effectiveness of hybrid ensemble learning in financial series forecasting, providing a more accurate and robust approach to dealing with outlier sacrificial data.
Keywords:	Time Series Analysis, Quantile Regression, Outlier Resilience, Machine Learning Framework, Robust Prediction.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	THE INFLUENCE OF SOCIAL MEDIA AND DIGITAL COMMUNICATION ON THE EVOLUTION OF VOCABULARY AND GRAMMATICAL STRUCTURES
Author:	VIKTORIIA SIKORSKA, OKSANA SNIGOVSKA, HANNA PEREDERII, ALINA АNDROSHCHUK, OLEKSANDR KALISHCHUK
Abstract:	The article discusses the impact of emerging communication technologies and social networks on the development of lexical and grammatical norms of the English language. The study is dedicated to the most important tendencies in language evolution, i.e., the emergence of neologisms, acronyms, abbreviations and borrowings, and grammatical simplifications and non-standard syntactic structures. Its importance is due to the need to investigate the mechanisms of language norm adaptation into the ever-changing digital environment, reshaping traditional language standards and communication methods. The research is based on the study of linguistic features of five popular sites (Twitter, Facebook, Instagram, TikTok, Reddit), which allows us to identify the specifics of the use of linguistic innovations in different situations of online communication. The article aims to determine the nature and causes of digital language changes, systematise their lexical and grammatical manifestations, and assess the impact of age and social factors on language dynamics. The study used a set of methods: content analysis, comparative and contrastive analysis, sociolinguistic approach, and descriptive analysis. The material was 250 text samples from five digital platforms. According to the research results, social networks are an effective mechanism for linguistic innovation, creating novel communication models and evolving forms of classical languages to digital ones. It has been established that different platforms have some linguistic features: Twitter is characterised by the active shortening of words and phrases, and TikTok and Instagram utilise non-standard grammatical forms with ironic or humorous connotations. Reddit is characterised by language play and violation of traditional syntactic rules. The research also revealed a strong dependency of language variations on users' age and social qualities: young people are the primary agents of language development. At the same time, their seniors keep traditional language norms. The findings may be used in future research on digital linguistics, namely how social media affects academic writing, professional jargon, and the long-term restructuring of the language system.
Keywords:	Social Networks, Language, Language Change, Online Communication, Communication, Language Tools
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	REVIEW ON PROGRAMMING LANGUAGE LEARNING MODELS AND INSTRUCTIONAL APPROACHES: CURRENT TRENDS AND FUTURE DIRECTIONS
Author:	SANAL KUMAR T S, R THANDEESWARAN
Abstract:	Programming has become a fundamental subject in the academic curriculum, but the learning process presents unique challenges. It requires not only systematic study and dedication but also the application of logical thinking and problem-solving skills in practical contexts. This complexity makes programming particularly difficult for beginners, who often face challenges in understanding foundational concepts like sequencing, decision-making, and looping. As a result, various pedagogical methods have evolved to address the difficulty in programming learning. To better understand the recent developments in programming educational approaches, we aim to provide a detailed review of learning models and instructional approaches in programming learning. Following this, we explore the cognitive factors influencing the learner and the essential aspects of the learner’s learning style and preferences in programming education. Finally, we conclude the review by discussing how these techniques can be combined to formulate future pedagogical approaches in programming instruction. Consequently, this review proposes integrating the learning style model with adaptive e-learning environments (ALE) in a blended learning approach as a better solution to address the hurdles of programming learning difficulty. Given this, the review paper provides a comprehensive overview of the programming learning environments, strategies, instructional approaches, cognitive factors, and learning styles leveraged in programming education, which future researchers can utilize.
Keywords:	Programming Education; Difficulty In Programming; Instructional Methods; Learning Style Models; Adaptive Learning Environments.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	VADER-RLA: A REINFORCEMENT LEARNING-AUGMENTED SENTIMENT ANALYSIS MODEL LEVERAGING VADER FOR CONTEXT-AWARE EMOTION CLASSIFICATION
Author:	KOLLI. SAI BHAVANA, DR. SENTHIL ATHITHAN, Dr. NESARANI ABRAHAM
Abstract:	VADER-RLA is presented as a hybrid architecture that utilizes reinforcement learning through VADER, expanding the contextual applicability of already established sentiment scoring in favour of comprehensive emotion classification, demonstrating that this framework can bridge the gap of ability to properly predict sentiment in tandem with trustworthiness. The most novel aspect of this work is the application of reinforcement learning to the problem of adaptive sentiment classification, hence allowing the model to perform real-time optimization of the sentiment scoring in changing linguistic surroundings. Proposed approach overcomes lexicon-based models and deep learning models, with respect to adaptability, precision and interpretability. Results show that VADER-RLA consistently exceeds traditional methods, yielding impressive accuracy and robustness and successfully detecting sophisticate sentiments like sarcastic and mixed ones. The experimental results indicate that VADER-RLA achieved an accuracy of 92.3%, an F1-score of 0.89, precision of 91.7%, and recall of 90.5%, demonstrating significant improvements over the baseline VADER model. These findings highlight the potential of VADER-RLA to provide a robust, adaptive solution for sentiment analysis in environments with rapidly changing linguistic trends.
Keywords:	Adaptability, Context-Aware, Emotion Classification, Hybrid Approach, Performance Metrics, Reinforcement Learning, Sentiment Analysis, VADER
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	EMPOWERING ONLINE SHOPPING SENTIMENT ANALYSIS USING TENACIOUS ARTIFICIAL BEE COLONY INSPIRED TAYLOR SERIES-BASED GAUSSIAN MIXTURE MODEL (TABC-TSGMM)
Author:	G.M. BALAJI, K. VADIVAZHAGAN
Abstract:	Sentiment Analysis has become increasingly important in online shopping, where consumers rely on reviews and feedback to make informed purchasing decisions. However, accurate sentiment classification poses challenges, such as handling nuanced language and varying review lengths. This study introduces a novel approach called the Tenacious Artificial Bee Colony inspired Taylor Series-based Gaussian Mixture Model (TABC-TSGMM) to address these challenges.TABC-TSGMM leverages two key components: the Taylor Series-based Gaussian Mixture Model (TSGMM) and the Tenacious Artificial Bee Colony (TABC). TSGMM captures complex sentiment patterns in text data, while TABC optimizes the model’s performance through an intelligent search strategy.TABC enhances TSGMM by optimizing model parameters, making the sentiment analysis process more robust and accurate. To evaluate the effectiveness of TABC-TSGMM, we conducted experiments on an Amazon product review dataset, focusing on electronic products. The results demonstrated superior classification accuracy, showcasing the potential of this approach to empower online shopping sentiment analysis.
Keywords:	Sentiment Analysis, ABC, GMM, Taylor Series, Local Search. Optimization
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text

Title:	CLOUD COMPUTING ADOPTION IN E-GOVERNMENT SERVICES: A COMPREHENSIVE POST-COVID-19 SYSTEMATIC REVIEW AND FUTURE DIRECTIONS
Author:	RAGHED ALKHASAWNEH, ZAIHISMA CHE COB, and ALIZA BINTI ABDUL LATIF
Abstract:	Cloud computing has emerged as a pivotal technology in transforming public sector services, particularly in the wake of the COVID-19 pandemic, which accelerated the global shift toward digital governance. This study conducts a comprehensive systematic literature review to examine the adoption of cloud computing in e-government services, underscoring its growing relevance in enhancing efficiency, scalability, and citizen engagement. The goal of this review is to identify the key factors influencing cloud adoption in government organizations, explore the most frequently applied IT/IS theoretical frameworks, and assess how the pandemic has reshaped adoption trends and priorities. Based on the analysis of 50 peer-reviewed studies from seven scholarly databases, the study hypothesizes that cloud adoption in e-government is significantly influenced not only by technological and organizational factors but also by external pressures such as public health crises and evolving citizen expectations. The findings confirm this hypothesis by revealing that while cloud computing offers substantial advantages, such as cost-effectiveness, improved accessibility, and service innovation, its adoption is challenged by data security concerns, legal barriers, and varying levels of institutional readiness. Additionally, the review identifies a shift toward citizen-centric service models post-pandemic, emphasizing the need for inclusive and resilient digital infrastructures. The study also highlights a gap in the literature, noting a limited diversity in publication sources and keywords, and encourages future research to broaden its scope. Practically, the insights gained can support policymakers and decision-makers in designing more adaptive, secure, and user-focused cloud-based e-government services. This review uniquely contributes by contextualizing cloud adoption within the post-COVID-19 digital transformation of the public sector.
Keywords:	Cloud Computing, E-government Services, Adoption, Benefits, Challenges, Systematic Literature Review, IS/IT Models, COVID-19.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2025 -- Vol. 103. No. 10-- 2025
Full Text