Diabetes is a non-communicable disease affecting people of all ages worldwide, therefore, early detection using machine learning techniques is crucial. This study aims to predict diabetes using multiple machine learning algorithms, performance metrics, and holdout validation on an Egyptian dataset. The dataset was divided into four age groups, including paediatric, early adulthood, middle age, and geriatric. Ten algorithms were applied and validated using 80:20, 70:30, and 60:40 split ratios with accuracy, precision, and recall as evaluation metrics. Results showed that Random Forest, Extra Trees, and Support Vector Machine performed best in the paediatric group, while Gradient Boosting, Random Forest, and Support Vector Machine achieved superior performance in early adulthood, middle age, and geriatric groups. In contrast, Decision Tree, K-Nearest Neighbors, and AdaBoost consistently demonstrated lower performance. Further analysis reveals that classification performance varies significantly across age groups, with the middle age and geriatric groups achieving the highest accuracy above 0.99, followed by the paediatric group 0.98–0.99, while early adulthood exhibits comparatively lower performance due to increased class overlap. Confusion matrix results indicate strong diagonal dominance in higher-performing groups, reflecting better class separability, whereas performance heatmaps confirm that top models maintain a balanced trade-off between accuracy, precision, and recall with minimal variation across different data splits. Feature importance analysis shows that higher performing models rely on a small number of dominant predictors, particularly in the middle age and geriatric groups, while more distributed feature contributions in early adulthood reduce predictive effectiveness. Therefore, the findings demonstrate that ensemble methods provide robust and consistent performance, and that age-based dataset segmentation enhances classification accuracy and model stability.

Ocular diseases such as cataracts, glaucoma, age-related macular degeneration, and diabetic retinopathy remain major contributors to global visual impairment and blindness, where early detection is critical for effective intervention. While color fundus photography is widely used for retinal screening, manual interpretation is time-consuming and prone to error, highlighting the need for automated, accurate, and interpretable diagnostic solutions. In this study, we propose a deep learning based framework for the automated classification of ocular diseases and normal cases using medical images. The framework incorporates a comprehensive preprocessing pipeline and leverages a scalable Apache-powered big data environment for efficient feature extraction and model training on large-scale datasets. A Convolutional Neural Network (CNN) was proposed and benchmarked against state-of-the-art architectures including VGG19, ResNet50, and GoogLeNet, achieving superior performance with an accuracy of 97%, precision of 93%, recall of 97%, and F1-score of 93%. To enhance interpretability and clinical trust, Gradient-weighted Class Activation Mapping (Grad-CAM) was integrated, generating heatmaps that highlight the most discriminative retinal regions influencing predictions. The proposed approach not only achieves high diagnostic accuracy but also ensures transparency, scalability, and clinical relevance, making it a promising step toward real-world deployment of explainable AI systems in ophthalmology and broader healthcare applications.

Bipolar disorder (BD) is still one of the most incapacitating of neuroaffective disorders in psychiatry. The strong mood swings from states of euphoria to depression often destabilize interpersonal relationships and can undo occupational stability. Early and reliable diagnosis facilitates prompt pharmacological intervention and mental-health education that may protect not only the patient and their immediate social circle but also the entire social structure from general distress. In this research study the performance of machine learning algorithms such as random forest (RF), support vector machine (SVM) and gradient boosting (GB) has been investigated for classification and prognostication of BD and its subtypes. The machine learning models were validated using a clinical dataset, which included 120 participants: 28 of BD I, 31 of BD II, 31 of Major Depressive Disorder and 30 healthy controls. Model performance was evaluated with stratified cross-validated train-test-split and a set of metrics, including accuracy, precision, recall, F1-score, and Receiver Operating Characteristic - Area under the Curve (ROC vs. AUC). In other words, the RF model had the highest accuracy (88%), precision (90%), and recall (88%). The discriminative performance of RF and SVM models was comparable with an ROC-AUC of 97\%. These results emphasize the potential of machine learning (ML), specifically ensemble techniques like Random Forest (RF), as an effective supplement to traditional early clinical diagnosis in bipolar disorders and related psychiatric illnesses

This systematic literature review (SLR) investigates search-based strategies for generating combinatorial test suites using covering arrays (CAs) to efficiently test system interactions. Conducted following PRISMA guidelines, the review analyzes 91 primary studies published between 2003 and 2025, selected through a rigorous process from major academic databases. The identified strategies are categorized into five types: standard, mix, adaptive, hybrid, and hyper-heuristic, based on their underlying algorithmic approaches, including swarm intelligence, evolutionary algorithms, and hyper-heuristic techniques. Each strategy is examined in depth, evaluating its effectiveness in generating high-quality combinatorial test suites. The review also highlights challenges in applying these strategies to varying software testing scenarios. Based on the findings, it provides practical insights to enhance their application and effectiveness in real-world contexts. This work supports broader adoption of search-based testing to improve software quality and reduce defect rates.

This study presents a comprehensive analysis of personality traits and learning skills among Technical and Vocational Education and Training (TVET) learners, leveraging advanced feature engineering and robust statistical methods. The primary objective is to investigate the influence of personality traits on learning outcomes and map these traits to specific skills across various TVET trades. Using a four-step methodology, we collected demographic data and Big Five Inventory (BFI) personality trait scores from 1,076 trainees across eleven Vocational Training Institutes (VTIs), specializing in eighteen distinct trades. Descriptive statistics, Spearman’s rank correlation analysis, and trade-specific BFI numerical scaling were employed to examine the relationship between personality traits and skill proficiency. Key findings reveal: (i) significant correlations between specific personality traits and trade proficiency, with notable variations across age and gender groups, and (ii) successful skill-to-trade mapping using BFI traits. These results highlight the potential of personality-aware approaches in vocational training to enhance occupational diversity and economic integration. By demonstrating the importance of aligning individual traits with skill development, this study contributes to the growing discourse on personalized vocational education. Future research should explore longitudinal studies to assess the long-term impact of personality traits on skill acquisition, career progression, and adaptability in vocational settings.

The demand for privacy-preserving machine learning has led to the rise of Federated Learning (FL), where multiple clients collaboratively train a model without sharing raw data. Despite its privacy benefits, FL is vulnerable to Byzantine failures, where malicious or faulty participants inject corrupted updates, threatening model integrity. To address this, a range of Byzantine-resilient aggregation techniques have been proposed, including statistical filters (e.g., Trimmed Mean, Krum), trust-based weighting, cryptographic protocols, and hybrid strategies. This paper presents a systematic literature review (SLR) of these defenses, evaluating their robustness, scalability, and suitability for real-world applications. Challenges such as non-IID data, adaptive attacks, and trade-offs between security and efficiency are critically examined. In addition, we explore emerging trends such as domain-specific defenses, energy-aware FL, quantum-resilient methods, and federated zero-knowledge proofs. A novel classification of hybrid approaches and a standardized benchmarking framework are proposed to guide future research. This review aims to support the development of resilient, efficient and scalable decentralized learning systems in adversarial environments.

Magnetic resonance imaging (MRI) serves as a crucial diagnostic tool, particularly for brain tumors where early detection significantly improves patient prognosis. The growing use of deep learning in medical imaging has led to substantial progress, yet the opaque nature of these models creates barriers to clinical acceptance, especially for critical applications such as tumor diagnosis. Our research applies explainable AI (XAI) techniques to improve the transparency of CNN-based brain tumor detection using MRI data. Working with a dataset containing 7,022 images spanning four tumor categories, our model attains 80\% accuracy while employing occlusion sensitivity analysis to produce visual interpretations. These heatmaps identify the most influential regions for predictions, giving clinicians insight into the model's decision process. This XAI integration enhances both understanding and accountability in healthcare AI systems, facilitating more reliable diagnostic tools.Precise early identification of brain tumors through MRI dramatically affects survival outcomes, though human interpretation remains time-consuming and variable. While CNNs show impressive classification results, their unclear reasoning limits clinical implementation. Our study introduces an XAI approach that pairs an accurate CNN classifier (80% on 7,024 multi-class scans) with occlusion analysis to create intuitive visual explanations. By methodically altering image areas and measuring prediction variations, we generate heatmaps that accurately pinpoint tumor-distinguishing features, matching radiological assessment. Comparative results demonstrate occlusion analysis's superiority over gradient methods like Grad-CAM in spatial precision for tumor classification (meningioma, glioma, pituitary). This research progresses clinically useful AI by connecting model effectiveness with interpretability in brain tumor imaging.

In this paper, we present the design and implementation of a smart water quality monitoring system for the Tharparkar region of Pakistan, where access to clean water is limited. The system utilizes Internet of Things (IoT) sensors and machine learning algorithms to assess and predict water quality. Parameters such as pH, turbidity, and total dissolved solids were continuously monitored using IoT sensors deployed in three strategically selected groundwater wells in Tharparkar. The collected data was transmitted wirelessly to a central server, where a Support Vector Regression model was applied to analyze water quality trends and classify samples as polluted or unpolluted. The results demonstrate the system's effectiveness in providing accurate, timely, and location-specific information, enabling early detection of contamination, and supporting proactive water resource management. This work highlights the potential of integrating IoT and artificial intelligence to address water scarcity and quality challenges in an underdeveloped region.

Software testing is integral to ensuring the functionality and quality of applications. This study highlights the implementation of Cypress and Node.js by the Nokia RON team to address the challenges of GUI testing. The adoption of Cypress enabled comprehensive and targeted testing, alongside efficient resolution of dependency and third-party package issues through selective installation. By integrating Cypress with Cucumber, an easy-to-use interface was developed to transform smoke test checklists into Gherkin syntax, enhancing readability and adaptability. Additionally, the use of Mochawesome Reporter provided detailed HTML reports, facilitating issue tracking and quick resolutions. This methodology, supported by a structured questionnaire, fostered stakeholder satisfaction and collaboration, resulting in an interactive and effective testing environment. The findings emphasize the role of Behavior Driven Development (BDD) in streamlining automated testing, improving communication among stakeholders, and ensuring higher software quality.

The utilization of machine learning (ML) methods has the potential to address the challenges posed by the rapid growth of student-related data, enabling better predictions of student performance and supporting informed managerial decisions. These techniques analyze data through advanced models and algorithms to forecast academic outcomes. This research focuses on identifying key factors that influence student performance using ML approaches. By leveraging statistical and classification algorithms, machine learning enhances the accuracy of predictions. The research explores relevant factors and applies in state-of-art models to achieve precise performance predictions. Various studies have employed ML techniques to predict student success, highlighting its broad applicability. This research proposes a framework for assessing students' academic achievements. The dataset includes information such as demographic details, prior academic records, and family background. Data was sourced from students across multiple universities using online surveys, comprising 24 attributes adapted from prior research. The objective is to identify the critical attributes that significantly affect student performance. It also evaluates distinguish classification techniques to enhance prediction accuracy. Experimental findings reveal that the Support Vector Machine (SVM) outperforms other methods, achieving a maximum accuracy of 62.50%. This research proposed the effective prediction tools can be developed to improve educational outcomes effectively.