Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
Find Your Way: An IOS-Based Travel Planning Application with Route Optimization Using Agile Methodology
No full textAims: This study aimed to develop Find Your Way, an iOS-based travel planning application that addressed user needs through the Agile development approach.
Study Design: The study applied Agile methodology, consisting of six iterative phase such as planning, design, development, testing, deployment, and review, executed over two development cycles to refine the application based on user feedback.
Place and Duration of Study: The research was conducted in Indonesia over a five-month period, from December 2024 to April 2025.
Methodology: Researchers conducted interviews with four participants aged 18–35 to gather functional requirements. The data were analyzed using thematic analysis to identify key user needs. These needs guided the design and implementation of the application using Swift and MVVM Clean Architecture. The core features developed included route optimization using Particle Swarm Optimization (PSO), destination recommendations, saving and managing travel routes, and customizable map settings. Testing was conducted using blackbox and whitebox methods. The app was deployed via TestFlight using CI/CD with Xcode Cloud. In the review phase, 30 users aged 18–25 completed a User Acceptance Test (UAT) to evaluate the app\u27s usability and performance.
Results: The study produced Find Your Way, a travel planning application tailored to user-identified functional needs. Five key features were implemented, including destination information, location and route recommendations, as well as storage and map display settings. The application achieved a UAT score of 97.33%, indicating a high level of user satisfaction. Testing confirmed the application\u27s functionality and program logic through blackbox and whitebox methods.
Conclusion: The Find Your Way application was successfully developed and met user functional requirements. The high UAT score of 97.33% demonstrated that the app effectively fulfilled its intended purpose, validating the Agile development approach used in this study
Hybrid AI-Human Architecture for Real-Time Customer Support: Leveraging Retrieval-Augmented Generation
No full textThis paper presents the design and implementation of a real-time, embeddable customer service system that integrates Retrieval-Augmented Generation (RAG) with human-in-the-loop (HITL) escalation. The system addresses common limitations in traditional customer support, such as slow response times and inconsistent service, by combining intelligent automation with timely human intervention. Built using FastAPI, ReactJS, PostgreSQL, and Centrifugo for WebSocket communication, the architecture supports low-latency interaction and seamless transitions between bot and human agents. At its core, the system leverages OpenAI’s GPT-3.5-turbo model to generate responses informed by user-provided domain-specific intent files.
To evaluate the effectiveness of the system, we conducted controlled testing using metrics such as average response time, session resolution rate, escalation latency, and accuracy of chatbot replies. Results showed that 71% of user queries were successfully handled by the chatbot alone, with an average response time under 3 seconds and human agent intervention occurring in less than 3 minutes for escalated cases. These outcomes suggest that the proposed system offers a scalable and intelligent alternative for real-time customer engagement, particularly in environments where domain-specific support and rapid escalation are essential
Signal Optimization and Energy-Efficient Design of Reconfigurable Intelligent Surface (RIS)-Assisted Non-Orthogonal Multiple Access (NOMA) Networks: A Simulation-Based Investigation
No full textThis study looks at signal optimization and energy-efficient design in Reconfigurable Intelligent Surface (RIS)-assisted Non-Orthogonal Multiple Access (NOMA) networks using simulation. RIS improves wireless communication by reflecting signals intelligently. This boosts spectral efficiency and cuts down on inter-user interference, which are major challenges in NOMA systems. The researchers developed a model based on Simulink to evaluate performance across different user distances. They compared traditional NOMA with RIS-assisted setups using various reflection methods. The results show that optimized RIS can boost signal strength by as much as 15 dB compared to non-RIS setups and by 7 to 8 dB compared to fixed-phase RIS. Spectral efficiency improved by 60 to 80%, rising from 2.5 bps/Hz to 4.6 bps/Hz. Energy consumption fell by about 30%, with power needs dropping from 130 mW to between 88 and 95 mW due to better signal direction. These improvements highlight the potential of RIS-assisted NOMA for better performance and energy savings. This makes it a good fit for next-generation wireless systems like 5G, 6G, IoT, and vehicular networks. The findings offer useful insights for configuring RIS effectively and allocating power, which supports scalable deployment in complicated communication settings
Adaptive Hybrid Data Preprocessing for Homogeneous Healthcare Data Integration and Ontology Construction
No full textHealthcare data, whether typically collected in a hospital environment (e.g., electronic health records, laboratory tests, etc.) or collated into an organized database for research and analytics, requires thorough preliminary examination in the form of data preprocessing to ensure trustworthy analysis and reliable semantic modeling. Inconsistent and heterogeneous data remain major obstacles in building effective ontologies, which are essential for semantic data integration. This paper presents an adaptive hybrid data preprocessing technique tailored for homogeneous data environments, aiming to enhance ontology construction. By integrating and customizing existing data cleaning methods, the approach dynamically addresses dataset-specific inconsistencies. AHPD is a modular pipeline that implements statistical, rule-based, and semantic-based methods and works to clean, normalize, and harmonize datasets typically structured, collected, or obtained from various component parts of a hospital. AHPD functions include dealing with missing data dynamically, maintaining awareness of inconsistencies, correcting inaccuracies, dealing with inter-dataset dependencies, and normalized schema alignment, resulting in data of reliable quality for analysis and semantic applications. From there, cleaned data files transformed into OWL-based ontologies can facilitate the inference and reasoning capabilities for intelligent querying. The performance of the ontology, enhanced by AHPD, was evaluated through the execution of SPARQL queries with high precision, recall, and F-measure, representing relevant clinical events and dependencies. The research concluded that AHPD improved data quality realized through analysis and compressed qualities of data, enabling practical construction of ontology and realistic potential of semantically informed smart applications to support integration of healthcare data and intelligent retrieval of health knowledge
Predictive Analysis of Student Engagement and Academic Performance in Virtual Learning Environments Using a Hybrid Markov: Machine Learning Model
No full textThe study aims at improving the prediction of student engagement and academic success in virtual learning environments (VLEs) by proposing a hybrid ensemble model that integrates Markov Chains, Hidden Markov Models (HMMs), and supervised machine learning algorithms. Traditional models often fail to capture the temporal dynamics of student behavior or provide timely and personalized interventions. To address these limitations, this study leverages sequential modeling and classification techniques on behavioral data collected from 500 students across multiple online courses.
The methodology includes data preprocessing, feature extraction, and model training using Decision Trees and Support Vector Machines (SVMs), alongside probabilistic modeling through Markov Chains and HMMs. Model evaluation was conducted using accuracy, F1-score, precision, recall, ROC-AUC, and confusion matrices.
Results show that the hybrid ensemble model outperforms individual models, achieving an accuracy of 91.9% and an F1-score of 89.9%. Forum participation and assignment completion emerged as the most influential predictors. Temporal modeling revealed that students in the high engagement state tend to remain consistent, while those in the low engagement state are unlikely to improve without support. Medium engagement students demonstrated the highest behavioral volatility, highlighting the importance of adaptive interventions.
The proposed model not only enhances predictive accuracy but also provides interpretable insights for early detection and support. These findings offer practical implications for educators, academic advisors, and learning system designers seeking to optimize student outcomes and retention in online learning contexts. Future research could explore integrating additional engagement indicators such as emotional and social interaction metrics, as well as testing the model across diverse cultural and institutional settings to further validate its generalizability
NetShield: A User-Centric Deep Learning Framework for Real-Time Network Anomaly Detection and Resolution
No full textThe growing dependence on digital networks has made network anomalies like performance degradation and critical security threats more disruptive. Crucially, existing detection systems are often too technical and lack practical guidance for average users. This study addresses that gap by developing a machine learning-based framework that detects a wide range of network anomalies and notifies users with simple, actionable solutions. The aim was to validate a hybrid deep learning model that combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to ensure effective detection. The model was trained and evaluated on five diverse benchmark datasets, enhancing robustness and generalizability. The results showed that the model achieved F1-scores above 97% across all datasets, outperforming traditional machine learning approaches. A fully functional prototype was developed to convert these outputs into real-time notifications, offering step-by-step guidance. For instance, upon detecting a phishing attempt, the system can automatically block the site and advise the user to never enter credentials on suspicious links. This research provides a validated, user-friendly framework that bridges the gap between technical anomaly detection and everyday cybersecurity practices, empowering users to take an active role in protecting their digital environments
An Integrated Economic Model of Investments in Mining Heaters: Balancing Energy Costs, Revenue Streams, and Climate Goals
No full textIt develops and applies one integrated economic-mathematical model of investment in mining heaters that yield heat together with the revenue of cryptocurrency. Methodologically, this article employs a four-block study: (i) benchmarking the energy efficiency of compared devices; (ii) scenario-based techno-economic modeling relating dynamics in electricity tariffs to network difficulty in cryptocurrencies and value for carbon credits; (iii) systematic CAPEX/OPEX accounting, and iv) evidence synthesis from long-term PPAs and free-/liquid-cooling case studies. Novelty accrues by bringing together in one framework the taxation of electricity-tariff dynamics, network-difficulty-volatility for cryptocurrencies, and carbon-credit-valuation-all factors that meet across four methodological blocks related-ASIC efficiency benchmarking; scenario-based profitability analysis; CAPEX/OPEX synthesis; and evidence from long-term PPAs and free-cooling case studies. Therefore, this paper brings an investment efficiency assessment unifier between the comparison of CAPEX with OPEX under different climatic and market contexts whereby three complementary revenue streams are specified computational as well as thermal and carbon. At place-representative residential/commercial tariffs mining-heaters may best offset a very great share of heating with digital incomes converting almost all consumed electricity into useful heat by the path. Free-cooling and liquid-cooling strategies in cold regions significantly lower capital and operating requirements-on the order of 60-80% data-center cooling normally makes further net positive project economics-improving even more so because being compared applies a liability rather than an amount to be set off. Computational, thermal, and carbon revenues mitigate their own price cycle and seasonal demand swings through diversification. It gives a practical decision model for private investors, district-heating operators, and DER stakeholders to size, price, and de-risk mining-heater deployments under rising tariffs and tightened decarbonization targets whereby such systems are both economically-and climate-advantageous
AI-Driven Inventory and Supply Chain Optimization: A Comparative Review of Tree-Based Machine Learning and Sequence Deep Learning Models
No full textArtificial intelligence (AI) is essential in supply chain and inventory management due to the increasing complexity, volatility, and unpredictability of these processes. Demand variations, seasonality, and nonlinear dependencies are frequently missed by conventional statistics and optimization techniques. Models for deep learning (DL) and machine learning (ML) have lately demonstrated significant promise for raising predicting accuracy, lowering costs, and enhancing robustness. Tree-based machine learning techniques like Random Forest, XGBoost, LightGBM, and CatBoost are prized for their effectiveness and interpretability when working with structured data. While they require more computing power, sequential deep learning models such as Temporal Convolution Networks (TCN) and Long Short-Term Memory (LSTM) are superior at simulating temporal dependencies and generating precise long-term predictions.
This study uses the PRISMA methodology to comprehensively analyze 42 peer-reviewed articles that were published between 2019 to August 2025. With the help of metrics like RMSE, MAE, and MAPE, the comparison assesses forecasting accuracy, scalability, interpretability, and processing overhead. The results show the trade-offs between the two methods and point out areas that need more research, such as the dearth of integrated comparative studies and the scant attention given to hybrid models. Amazon, Walmart, and Alibaba real-world examples demonstrate the usefulness of these techniques.
These results highlight the potential of explainable AI and hybrid architectures to bridge the gap between theory and practice. For scholars and industry professionals looking for scalable, data-driven, and robust supply chain solutions, this evaluation offers insightful information
Data Sovereignty and Digital Health Transformation in Saudi Arabia: A Review with Policy Implications for Vision 2030
No full textThe rapid expansion of artificial intelligence (AI) and digital health technologies has made data sovereignty a critical policy concern. In Saudi Arabia, Vision 2030 emphasizes technological innovation, economic diversification, and resilient digital infrastructures, positioning data governance as a foundation for national transformation. This review examines how data sovereignty strategies can support the Kingdom’s digital health transformation while ensuring compliance with international standards. Using a qualitative approach, the study synthesizes policy documents, institutional reports, and peer-reviewed literature. The analysis identifies both enablers and barriers: strong institutions such as the Saudi Data and Artificial Intelligence Authority (SDAIA) and the National Data Management Office (NDMO), alignment of the Personal Data Protection Law (PDPL) with frameworks like the GDPR, but also challenges including rigid intellectual property rules, fragmented cloud data governance, and weak sector-specific policies. The study’s objectives are to clarify these challenges, assess their impact on generative AI adoption, and recommend reforms that align with Vision 2030’s innovation goals. Policy recommendations include adaptive IP frameworks, sector-specific governance models, inter-agency coordination, and public–private partnerships to expand access to lawful datasets. The findings demonstrate that data sovereignty, when strategically designed, functions not as a barrier but as an enabler of secure, scalable, and ethically responsible innovation in healthcare and beyond. This work contributes actionable insights for policymakers, regulators, and researchers in Saudi Arabia and offers lessons for other nations navigating the balance between sovereignty and competitiveness in the digital era
Behavioral Biometrics-Powered Continuous Authentication for Zero-trust Remote Work Environments: A Multi-factor Identity Verification Framework
No full textThis research presents a behavioral biometrics-powered continuous authentication framework designed for zero-trust remote work environments in healthcare. The study integrated keystroke dynamics, mouse movement patterns, and contextual risk factors into a multimodal system enabling seamless, real-time identity verification. Using a fused dataset of 3,600 samples from 24 users derived from public keystroke and mouse-dynamics repositories and augmented with healthcare-specific synthetic data where 79 behavioral features were engineered and normalized. Six machine learning and deep learning models were trained, including Random Forest, XGBoost, Support Vector Machine (SVM), LSTM, CNN-LSTM, and CNN, with Random Forest and XGBoost achieving the best performance at 98.25% accuracy, 0% Equal Error Rate (EER), and an AUC-ROC of 1.0000. The framework operated frictionlessly, with inference times below 2.6 ms, ensuring zero disruption to clinical workflows. Dynamic trust scoring enabled adaptive access control, while attack simulations across six threat scenarios yielded a 90.3% detection rate, including 100% for insider threats and zero-effort impersonation. Full compliance with HIPAA standards was validated through continuous monitoring, audit logging, and real-time threat response. The system outperformed traditional authentication methods in accuracy, usability, and security resilience. Despite strong results, limitations include constrained user diversity and simulated environments. The framework advances zero-trust principles by providing passive, high-precision authentication suitable for distributed healthcare systems. Future work should focus on longitudinal field deployment and adaptive modeling to address behavioral drift and emerging threats