Journal of Innovations in Computer Science and Engineering (JICSE)

Abstract The Social Internet of Things (SIoT) enhances device interoperability through social relationships but introduces significant trust management challenges in dynamic, resource-constrained environments. Existing machine learning-based trust models often rely on static training data or centralized architectures, limiting their adaptability and scalability in real-world SIoT deployments. This paper proposes GossipTrust, a novel decentralized trust model that integrates fog computing with adaptive control mechanisms for robust trust management. Our approach deploys machine learning models on fog nodes to enable distributed trust computation while minimizing resource constraints on individual devices. The core innovation combines a MAPE-K (Monitor, Analyze, Plan, Execute, and Knowledge) control loop for real-time attack detection with a gossip learning protocol that allows fog nodes to collaboratively refine trust models through peer-to-peer updates without central coordination. Experimental results demonstrate that GossipTrust significantly outperforms baseline models, achieving 13% higher accuracy, 14% higher success rate, 10% higher F-measure, and 12% lower loss across various trust attack scenarios. The proposed solution effectively addresses key SIoT challenges, including scalability, adaptability, and resource efficiency.

Abstract This study presents an applied, multi-stage machine-learning pipeline for maternal health risk classification using the Maternal Health Risk dataset (1,014 samples, six clinical features, and three risk classes). Six classifiers were evaluated across four preprocessing stages: training-fold standardization alone, Edited Nearest Neighbors (ENN)-based cleaning, Synthetic Minority Oversampling Technique (SMOTE)-based class balancing, and sequential ENN-SMOTE preprocessing. Within each fold of 10-fold cross-validation, the scaler, ENN, SMOTE, model fitting, and ensemble construction were performed using only the training fold; the corresponding test fold remained untouched until evaluation. A fixed majority-voting ensemble comprising Random Forest, Extra Trees, and XGBoost was evaluated using class-wise, macro-averaged, and weighted-averaged accuracy, precision, and F1-score. In Stage 4, the hybrid ensemble achieved a weighted accuracy of 91.5% ± 3%, precision of 88% ± 3%, and F1-score of 87.2% ± 3%. The ensemble's precision for the high-risk class exceeded 90% across all four stages. These findings indicate competitive predictive performance on this dataset; however, direct comparison with previous studies is limited by differences in preprocessing methods, validation protocols, and model configurations. External and prospective validation is required before clinical use.

Abstract Space–Air–Ground–Ocean Integrated Networks are being investigated as a communication framework for exchanging data across terrestrial, aerial, space, maritime-surface, and underwater environments. Environmental monitoring and resilient operation along strategic waterways expose limitations in conventional single-domain architectures, particularly when links differ sharply in propagation, capacity, delay, availability, and security exposure. This review assesses emerging communication infrastructures across the four domains and examines signal attenuation underwater, satellite-link delay, channel uncertainty, intelligent surface gateways, heterogeneous radio-frequency (RF), acoustic, and optical media, software-defined networking (SDN), network function virtualization (NFV), and adaptive protocols. Building on prior surveys of individual domains or themes, it jointly analyzes cross-domain architecture, heterogeneous transmission media, protocol convergence through intelligent surface gateways, SDN/NFV- and artificial intelligence (AI)-assisted network management, cyber resilience, and open research directions for strategically important maritime environments. The review concludes by identifying evidence gaps and validation priorities for interoperable, energy-aware, secure, and disruption-tolerant services spanning remote ocean regions and other heterogeneous environments.

Abstract Vision-Language-Action (VLA) models equipped with billion-parameter architectures offer unprecedented reasoning capabilities for embodied agents. However, their direct deployment in continuous control systems faces critical challenges. These challenges include high inference latency, kinematic instability, and severe fragility to visual perturbations. In this paper, we propose an uncertainty-aware hierarchical control architecture alongside a comprehensive evaluation framework. Our primary contributions are threefold: (1) an information-theoretic meta-controller that maps the VLA's predictive uncertainty heuristics (Shannon entropy and logit variance) directly to the planner's update rate and control policy; (2) a POMDP-grounded formalization of Cognitive Deadlock, which we resolve using an Extended Kalman Filter (EKF) as a Kinematic Spatial Memory; and (3) a Control-Coupled Chain-of-Thought (CoT) system where spatial reasoning actively constrains an Artificial Potential Field (APF). We validate our approach through rigorous statistical analysis across 15 independent simulation seeds. The empirical results demonstrate that direct End to End (E2E) deployment of State Of The Art (SOTA) VLAs frequently fails due to cognitive deadlocks. In contrast, our entropy-guided APF architecture maintains empirically bounded tracking behavior under severe visual disturbances. It successfully restores continuous kinematic actuation, completely overcoming the physical paralysis observed in direct policy execution.

Abstract The need for hospitals and care facilities for society is expanding as a result of the aging population. Furthermore, due to the limited hospital resources, it becomes more important to have a system for managing these resources. One of the technologies available for remote health monitoring is the Wireless Body Area Networks (WBANS). By employing this technology, unnecessary hospitalization can be prevented, and it is also a beneficial way to manage resource constraints. However, energy limitations due to battery storage constraints are challenging in WBANs. In many of these networks, it is not possible to recharge or replace the batteries. Thus, energy saving and extending the network’s life span are required. An effective method that has been suggested to reduce energy usage is rate-management-based sampling.

Materials and Methods: To reduce pointless communications, we provide an effective sampling rate method in this paper. Along with using Pulse Code Modulation (PCM) for obtaining digital signals and making the PAM signals appropriate for transmission, we used the Flat-Top Pulse Amplitude Modulated to minimize data transmission.

Results: The results illustrate that energy usage was significantly reduced. Using the PCM method resulted in a reduction in data transmission by up to 92.7%, and an improvement in energy consumption in the wireless network by up to 88% compared to baseline and NIS and SRCM.

Conclusion: Our aim was to reduce energy consumption by governing the rate of data sampling. By implementing our method, we fulfilled this goal.

Abstract The widespread adoption of Continuous Integration and Continuous Delivery (CI/CD) pipelines, particularly through platforms such as GitHub Actions, has introduced new challenges in ensuring pipeline reliability and predictability. Unexpected workflow failures can disrupt release cycles and increase operational costs. This paper presents a time-aware, data-driven approach for predicting CI/CD workflow execution outcomes. Historical failure patterns are extracted using time-aware cumulative aggregation. Feature engineering is performed incrementally, including Bayesian smoothing to stabilize sparse statistics, Last-K features (k = 5, 10) to capture short-term behavior, and interaction features to model nonlinear dependencies between entities such as repositories and workflow paths, forming the core feature-set (Base_Feature_Set). Additionally, textual features are extracted from workflow logs using regular expressions, capturing error presence, frequency, and message patterns, resulting in an enriched feature-set (Log-Enriched_Feature_Set). Using the large-scale GHALogs dataset, comprising over 513,000 workflow executions from approximately 25,000 repositories, we evaluate five models (Logistic Regression, Random Forest, HistGradientBoosting, XGBoost and CatBoost) under a consistent chronological split for both feature sets. Experimental results show that time-aware structured features alone yield moderate discriminative power (AUC 0.603–0.646). Adding log-derived features, which become available only after execution, raises the AUC of the final CatBoost model to 0.843. Because the log features describe the run being predicted, this improvement reflects post-execution failure-detection performance rather than pre-execution forecasting.

Abstract Pharmaceutical ecosystems involve interdependent enterprises, regulators, healthcare providers, patients, and digital platforms whose interactions create substantial organizational, governance, and interoperability challenges. Existing contributions address important parts of this problem through domain ontologies, enterprise-architecture frameworks, process models, and digital-platform perspectives, but remain fragmented at the ecosystem level. This study presents CM-PharmE 1.0, a business-architecture-informed and UFO-grounded conceptual model expressed in OntoUML. A PRISMA-guided review identified 380 records, of which 17 studies were retained for qualitative thematic synthesis. The resulting modeling concerns were organized into five interrelated domains and translated into UFO/OntoUML constructs through an explicit concept- and relation-selection procedure. Assessment combined a scenario-based comparative benchmark, qualitative review by four experts, a focused manual inspection of eight selected OntoUML-informed anti-pattern categories, competency-question analysis, illustrative vaccine-distribution instantiation, and repository-supported computational checks. All eight competency questions were operationalized as executable queries and returned the predefined bounded results against the formalized model and machine-readable scenario; automated ROBOT/HermiT reasoning also completed without reporting a logical inconsistency or unsatisfiable named classes in the current axiom set. The resulting evidence provides preliminary support for the model’s conceptual traceability, cross-domain coherence, ontological discipline, and bounded computational consistency, while reference-architecture adoption, contextual extensibility, operational interoperability, empirical completeness, and implementation effectiveness remain to be evaluated. CM-PharmE 1.0 is therefore positioned as an extensible conceptual baseline for governance reasoning, interoperability-oriented analysis, requirements engineering, and subsequent architectural development in regulated pharmaceutical ecosystems rather than as a complete or operationally validated ontology.

Abstract Context: Mobile application functional features are implemented through GUI event handlers, whereas user-visible behaviors emerge from meaningful sequences of these handlers. Therefore, ensuring correctness and detecting functional defects, especially non-crashing ones, requires test cases enriched with app-specific oracles to validate both individual handlers and higher-level workflows, called targeted test cases. While numerous GUI testing approaches have significantly advanced automated coverage and defect detection, systematically generating targeted test cases to verify the app-specific behaviors of mobile applications remains a challenge.

Objective: This study aims to address this gap by systematically generating targeted test cases to exercise both individual functional features and their meaningful, high-level user workflows in mobile applications.

Method: We propose a feature-oriented, model-based GUI testing approach, which represents the System Under Test (SUT) by introducing two complementary models: the Event Model (EM), which captures individual functional GUI event handlers and their parent–child dependencies, and the Event Relation Model (ERM), which captures meaningful sequences of handlers as user-level functional scenarios. Test cases generated from these models are enriched with GUI-level invariants that serve as app-specific oracles for runtime validation. We evaluate test effectiveness using coverage criteria and mutation testing.

Results: An empirical evaluation on several open-source mobile applications shows that the generated test cases achieve, on average, 85% statement coverage and a 97% mutation score. Moreover, GUI-level invariants improve fault-detection effectiveness by 19% to 32%.

Conclusion: The results demonstrate that combining feature-oriented behavioral modeling with app-specific test oracles, improve the effectiveness of the proposed approach over state-of-the-art testing approaches.

Abstract Software testing is a crucial and resource-intensive phase in the software development life cycle, where test data generation (TDG) remains a major bottleneck. While automated tools exist, human intuition is still valuable for resolving complex logic. To optimize this tedious task, we propose a 3D Game with a Purpose (GWAP) that crowdsources TDG. Unlike prior studies, our 3D environment maps program execution paths to interactive spatial puzzles and visually represents complex data structures, such as arrays and records, through customizable character attributes. This specific visual mapping enhances player interaction and addresses previous structural limitations. We evaluated the proposed approach using standard benchmark programs. Quantitative results demonstrate that players achieved 100% coverage across all 279 feasible execution paths—out of 511 total analyzed paths, 232 of which were infeasible—in an average of 29 minutes, whereas the baseline game failed to generate puzzles for 30.5% of these feasible paths (85 of 279) due to structural limitations. Furthermore, user evaluations show engagement among players with varying levels of expertise. The results confirm that the proposed 3D GWAP successfully handles complex input parameters while transforming a technical testing task into an effective, user-friendly environment.