-
Research Article
Improving Anomaly Detection Accuracy Using Fuzzy Cuckoo-Inspired Clustering and Optimization Techniques
Issue:
Volume 11, Issue 2, December 2026
Pages:
63-75
Received:
16 February 2026
Accepted:
1 June 2026
Published:
22 July 2026
Abstract: Anomaly detection is a critical task for identifying unusual patterns that may indicate security breaches, fraudulent transactions, or system failures in various application domains. Conventional anomaly detection techniques often experience high false positive rates and limited adaptability when handling complex, uncertain, or evolving datasets. To overcome these limitations, this study proposes a novel Fuzzy Cuckoo-Based Clustering Technique (F-CBCT) that integrates fuzzy logic with cuckoo search-based clustering and optimization. The proposed framework employs a decision tree classifier enhanced with fuzzy membership functions, enabling effective management of uncertainty during classification. Model parameters are optimized using a hybrid strategy based on Mean Square Error (MSE) and the Silhouette Index, improving clustering quality and classification accuracy. Experimental evaluations conducted on benchmark datasets demonstrate the effectiveness of the proposed approach, achieving a 96.86% detection rate, 97.77% accuracy, a 1.297% false positive rate, and an F-measure of 98.30%. Comparative analysis with existing state-of-the-art anomaly detection methods confirms that F-CBCT consistently outperforms conventional approaches in terms of detection capability, robustness, and reliability. The proposed technique effectively reduces false alarms while maintaining high detection performance, making it a promising solution for real-world anomaly detection applications across diverse and dynamic environments.
Abstract: Anomaly detection is a critical task for identifying unusual patterns that may indicate security breaches, fraudulent transactions, or system failures in various application domains. Conventional anomaly detection techniques often experience high false positive rates and limited adaptability when handling complex, uncertain, or evolving datasets. T...
Show More
-
Research Article
Investigation on Machine Learning Models for Predicting Diabetes Risk in Indian Populations
Gajendra Singh*
Issue:
Volume 11, Issue 2, December 2026
Pages:
76-84
Received:
8 July 2026
Accepted:
6 August 2026
Published:
2 September 2026
Abstract: Background: Diabetes mellitus is a major public health concern in India, with increasing prevalence driven by demographic, metabolic, behavioral, and lifestyle-related factors. Early identification of individuals at high risk of diabetes can support timely prevention and improve health outcomes. Machine learning (ML) approaches offer opportunities to identify complex and nonlinear relationships among multiple risk factors and may complement conventional statistical approaches. Objective: This study aimed to investigate and compare the performance of different ML models for predicting diabetes risk, with particular emphasis on identifying the most influential predictors and assessing the potential applicability of ML-based approaches for early risk stratification in Indian populations. Methods: A cross-sectional analytical approach was used to evaluate demographic, physiological, and lifestyle-related variables associated with diabetes risk. The study considered variables including age, body mass index (BMI), blood glucose, blood pressure, insulin, family history of diabetes, and physical activity. Data preprocessing included missing-value management, outlier identification, categorical encoding, and feature standardization. The dataset was divided into training (70%) and testing (30%) subsets using stratified sampling. Seven supervised ML algorithms—Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, K-Nearest Neighbors, Extreme Gradient Boosting (XGBoost), and Artificial Neural Network—were compared. Model performance was assessed using accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC). Pearson correlation, one-way ANOVA, and multivariate logistic regression were additionally used to examine associations between predictors and diabetes risk. Results: XGBoost demonstrated the strongest overall predictive performance, achieving an accuracy of 89.2%, precision of 0.88, recall of 0.87, F1-score of 0.87, and AUC of 0.93. Random Forest and the neural network also demonstrated strong performance, with accuracy of 87.6% and 88.5% and AUCs of 0.91 and 0.92, respectively. Glucose level was the most influential predictor, followed by BMI and age. Statistical analyses further supported the importance of these metabolic and demographic factors in diabetes risk prediction. Conclusion: Ensemble and advanced ML approaches, particularly XGBoost and Random Forest, demonstrated promising performance for diabetes risk prediction. Integrating ML with statistical analysis and interpretable AI approaches may strengthen early risk identification and support evidence-based diabetes prevention and personalized healthcare strategies in Indian populations. Further validation using larger, diverse, and multi center datasets is required before clinical implementation.
Abstract: Background: Diabetes mellitus is a major public health concern in India, with increasing prevalence driven by demographic, metabolic, behavioral, and lifestyle-related factors. Early identification of individuals at high risk of diabetes can support timely prevention and improve health outcomes. Machine learning (ML) approaches offer opportunities ...
Show More
-
Research Article
Explainable Sequence-Aware Deep Learning Framework for Potential Zero-Day Attack Detection and Cross-Domain Generalization in Enterprise Network Intrusion Detection
Issue:
Volume 11, Issue 2, December 2026
Pages:
85-111
Received:
9 September 2026
Accepted:
21 September 2026
Published:
30 September 2026
Abstract: Zero-day attacks remain a significant challenge in enterprise network security because their previously unseen characteristics can reduce the effectiveness of conventional signature-based intrusion detection systems. Although machine learning and deep learning have improved intrusion detection, many existing approaches are evaluated within a single dataset and often treat network traffic records as independent observations, providing limited evidence of temporal behavior and cross-domain generalization. This study proposes an Explainable Sequence-Aware Deep Learning Framework for Potential Zero-Day Attack Detection and Cross-Domain Generalization in Enterprise Network Intrusion Detection. The framework represents network traffic as overlapping sequences of 20 consecutive network-flow records and combines a one-dimensional convolutional neural network (1D-CNN), Bidirectional Long Short-Term Memory (Bi-LSTM), and four-head Multi-Head Self-Attention to learn local traffic characteristics, temporal dependencies, and informative relationships within network behavior. A shared representation supports both binary intrusion detection and multiclass attack classification, while SHAP and LIME provide global and local explanations of model decisions. CICIDS2017 serves as the source domain for model development, whereas UNSW-NB15 is maintained as an independent target domain for cross-domain evaluation. A stratified sample of 50,000 records is independently selected from each dataset, with SMOTE applied only to the CICIDS2017 training data. On the CICIDS2017 internal test set, the framework achieved 96.89% accuracy, 97.45% precision, 89.65% recall, 93.38% F1-score, 0.9961 ROC-AUC, and 0.9379 MCC for binary detection, while multiclass classification achieved 97.0% accuracy and 96.8% F1-score. On the independent UNSW-NB15 test set, binary detection achieved 87.70% accuracy and 90.56% F1-score, while multiclass detection achieved 80.35% accuracy and 59.57% F1-score. The findings demonstrate strong in-domain learning and useful cross-domain detection capability without retraining or fine-tuning. The cross-domain results also revealed the difficulty of transferring learned representations across different network environments. In this study, cross-domain evaluation is used to assess potential zero-day detection capability rather than to claim detection of a specifically verified zero-day attack.
Abstract: Zero-day attacks remain a significant challenge in enterprise network security because their previously unseen characteristics can reduce the effectiveness of conventional signature-based intrusion detection systems. Although machine learning and deep learning have improved intrusion detection, many existing approaches are evaluated within a single...
Show More
-
Research Article
ERGA-Phish: Optimized Ensemble Learning Framework with Ridge Regression and Genetic Algorithm for Improved Phishing URL Detection
Issue:
Volume 11, Issue 2, December 2026
Pages:
112-127
Received:
29 August 2026
Accepted:
11 September 2026
Published:
30 September 2026
Abstract: Phishing remains a persistent cybersecurity threat that exploits deceptive web links to compromise users and organizations. Although machine learning and deep learning techniques have demonstrated strong phishing detection capabilities, challenges related to class imbalance, feature redundancy, generalization, and model optimization remain. This study proposes an optimized stacking ensemble framework that integrates Decision Tree, Random Forest, Naïve Bayes, XGBoost, and Artificial Neural Network (ANN) classifiers, with an ANN serving as the meta-learner. The PhiUSIIL Phishing URL Dataset, obtained from the UCI Machine Learning Repository, contains 235,795 labeled URLs, comprising 134,850 legitimate and 100,945 phishing URLs. Data preprocessing involved duplicate removal, mean imputation, one-hot encoding, normalization, and Synthetic Minority Oversampling Technique (SMOTE), producing a balanced dataset of 269,700 instances. Ridge Regression was then applied to reduce the feature space from 57 to 50 relevant features. The stacking ensemble was further optimized using a Genetic Algorithm (GA) to identify suitable parameter configurations for improved predictive performance. Using an 80:20 train-test split, the proposed model achieved 99.54% accuracy, 98.75% precision, 98.85% recall, 98.68% F1-score, and 0.98 ROC-AUC. The proposed model outperformed the evaluated baseline classifiers and most benchmarked studies, although one reported approach achieved 99.89% accuracy. The findings demonstrate that combining heterogeneous classifiers, feature selection, class balancing, and evolutionary optimization can provide a strong framework for phishing URL detection.
Abstract: Phishing remains a persistent cybersecurity threat that exploits deceptive web links to compromise users and organizations. Although machine learning and deep learning techniques have demonstrated strong phishing detection capabilities, challenges related to class imbalance, feature redundancy, generalization, and model optimization remain. This st...
Show More
-
Research Article
Epistemic Injustice in Algorithmic Systems: A Frickerian Analysis of AI-Driven Knowledge Production
Joel Sadiki Charo*,
Kennedy Ondimu
Issue:
Volume 11, Issue 2, December 2026
Pages:
128-136
Received:
19 September 2026
Accepted:
4 October 2026
Published:
29 October 2026
Abstract: Artificial intelligence is playing a great role in the production, classification, retrieval, and communication of knowledge. Tools like large language models, recommender systems, automated decision-making systems, and search engines do more than just process information; they shape whose testimony is heard, which concepts become available for interpreting experience, and which claims are treated as credible or important. This paper argues that Miranda Fricker's theory of epistemic injustice offers a strong framework to ethically identify shortcomings in AI-enabled knowledge production; however, it also suggests that, the distinction made in classical accounts between testimonial and hermeneutical injustices needs to be extended in order to address them in the context of sociotechnical systems. Testimonial injustice can arise when algorithmic models reproduce credibility deficits through biased training data, proxy variables, ranking practices, moderation systems, or automated classification. On the other hand, hermeneutical injustice can emerge when certain communities are poorly represented in the conceptual and linguistic tools that AI systems use to create explanations, categories, and interpretations. The paper further argues that the opaque nature of many AI systems intensifies these harms because affected persons may be unable to understand, contest, or correct the epistemic processes that affect them. Building on Fricker, Floridi, Dotson, Medina, and contemporary work in AI ethics, the paper develops an epistemic justice framework centred on adequacy in represantation, credibility protection, interpretive inclusion, contestability, provenance, and institutional accountability. The proposed framework treats epistemic justice not merely as a desirable output property but as a design and governance requirement for AI systems that participate in consequential knowledge practices.
Abstract: Artificial intelligence is playing a great role in the production, classification, retrieval, and communication of knowledge. Tools like large language models, recommender systems, automated decision-making systems, and search engines do more than just process information; they shape whose testimony is heard, which concepts become available for int...
Show More
-
Research Article
Adaptive Artificial Intelligence Models for Enhancing Digital Inclusion and Sustainable Development in Underserved Communities
Joel Sadiki Charo*,
Kennedy Ochilo Hadullo
Issue:
Volume 11, Issue 2, December 2026
Pages:
137-142
Received:
21 September 2026
Accepted:
8 October 2026
Published:
29 October 2026
Abstract: Technology is playing a vital role across various sectors like education, healthcare, agriculture, financial services, employment, governance, and social development. However, we still see significant digital inequalities, especially in underserved communities, particularly in rural and low-income areas. Here, challenges such as limited connectivity, high technology costs, lack of digital skills, language barriers, and inaccessible digital services hinder meaningful participation in the digital economy. This conceptual paper delves into how adaptive Artificial Intelligence (AI) models can boost digital inclusion and support sustainable development in these communities. It brings together insights from literature on artificial intelligence, digital inclusion, adaptive learning, responsible AI, and sustainable development. The paper introduces the Adaptive AI Digital Inclusion and Sustainable Development (AIDISD) framework, which is built around five key dimensions: accessibility, personalization, contextual adaptation, resource efficiency, and community empowerment. It explores potential applications in education, healthcare, agriculture, financial inclusion, employment, and public services. Additionally, the paper highlights challenges related to infrastructure, data scarcity, algorithmic bias, privacy, cybersecurity, affordability, digital literacy, and institutional capacity. The analysis emphasizes that adaptive AI should be viewed as part of a larger socio-technical strategy rather than just a standalone tech solution. For inclusive deployment, we need local participation, suitable infrastructure, responsible data governance, human oversight, and ongoing evaluation. Ultimately, the paper concludes that well-designed adaptive AI systems can foster more inclusive and sustainable digital ecosystems, provided they are affordable, locally relevant, accessible, transparent, and aligned with the Sustainable Development Goals.
Abstract: Technology is playing a vital role across various sectors like education, healthcare, agriculture, financial services, employment, governance, and social development. However, we still see significant digital inequalities, especially in underserved communities, particularly in rural and low-income areas. Here, challenges such as limited connectivit...
Show More
-
Research Article
Ict-enabled Environmental Sustainability in Coastal Communities: Assessing Digital Governance and Blue Economy Development in Mombasa County, Kenya
Joel Sadiki Charo*,
Mvurya Mgala
Issue:
Volume 11, Issue 2, December 2026
Pages:
143-148
Received:
5 October 2026
Accepted:
16 October 2026
Published:
29 October 2026
Abstract: Coastal communities face increasing environmental challenges; including habitat degradation, pollution, and climate-related shocks that threaten livelihoods built around the ocean economy. Information and Communication Technologies (ICTs) are increasingly promoted as instruments of environmental governance, yet the extent to which digital governance capacity translates into sustainable blue economy outcomes in African coastal counties remains under-researched. This paper investigates the relationship between ICT-enabled digital governance and environmental sustainability in the context of blue economy development in Mombasa County, Kenya. Grounded in socio-technical systems theory and the public value of information systems perspective, the study adopts an embedded single-case study design combining a cross-sectional household and business survey, key-informant interviews with government and research institutions (NEMA, Mombasa County Government, KMFRI, KWS), and a documentary review of county digital-governance initiatives. The study proposes a Digital Governance Readiness Index (DGRI) integrating four dimensions digital infrastructure, institutional capacity, stakeholder participation, and data-driven decision making, and links these to blue economy outcomes through a conceptual model with six testable propositions. Expected contributions include an evidence-based assessment framework that county governments can use to diagnose governance gaps, a set of policy recommendations for aligning national blue economy strategy with grassroots digital inclusion, and a replicable methodology for other coastal counties. The paper indicates that ICT-enabled sustainability is contingent not merely on technology deployment but on participatory governance structures, inter-agency data sharing, and community digital literacy.
Abstract: Coastal communities face increasing environmental challenges; including habitat degradation, pollution, and climate-related shocks that threaten livelihoods built around the ocean economy. Information and Communication Technologies (ICTs) are increasingly promoted as instruments of environmental governance, yet the extent to which digital governanc...
Show More
