Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
Generative AI in Supply Chain: A Systematic Review of Opportunities, Benefits, and Challenges
No full textThis systematic review investigates the impact of Generative Artificial Intelligence (GenAI) on supply chain management (SCM). Though there is an acknowledgment of GenAI’s potential in SCM, there remains a lack of research that thoroughly examines the opportunities, tangible advantages, and challenges associated with GenAI adoption in SCM. This literature review seeks to analyze the role of GenAI in SCM and fulfill three primary objectives. First, it identifies the key opportunities for GenAI implementation in SCM, encompassing its applications in demand forecasting, inventory control, and decision-making support systems. Second, it assesses the tangible benefits observed from current implementations, such as cost savings and enhanced operational efficiency. Third, the study investigates the challenges and obstacles to effective GenAI adoption, focusing on technical limitations, data quality concerns, workforce skill deficiencies, and ethical issues. This research followed the systematic review methodology, reviewing and synthesizing information from 24 peer-reviewed articles published between 2023 to 2024. The findings reveal substantial potential for GenAI in SCM areas like demand forecasting, inventory optimization, and risk management while emphasizing significant data quality, integration, and organizational preparedness challenges. The study offers two practical implications: a) the findings provide actionable insights for supply chain professionals, highlighting how GenAI can be integrated into SCM; b) the research underscores the necessity for strong data governance, ethical AI usage policies, and adherence to evolving regulations.
Objective: This study aims to understand GenAI’s role in SCM, identify key opportunities for implementing GenAI in supply chain processes, analyze the tangible benefits reported in existing implementations, and examine the challenges and barriers to successful GenAI adoption.
Method: This review adhered to PRISMA guidelines, applying rigorous inclusion and exclusion criteria across multiple databases to yield 24 peer-reviewed articles from 2023–2024. Data extraction, quality assessment, and synthesis were conducted systematically to ensure unbiased insights into GenAI’s impact on supply chain management.
Results and Discussion: The results reveal that GenAI significantly enhances SCM processes such as demand forecasting, inventory optimization, and risk management, while also exposing challenges in data quality, system integration, and organizational readiness. The discussion contextualizes these findings within the theoretical framework, emphasizing practical implications and addressing study discrepancies and limitations.
Research Implications: This research expands the exploration of GenAI in SCM and encourages further investigation into emerging technologies to optimize supply chains. For the SCM industry, it highlights a strategic approach to GenAI adoption, emphasizing technology investments, workforce training, data integrity, and ethical practices to boost operational efficiency and resilience.
Originality/Value: This study contributes to the literature by examining the practical integration of GenAI into supply chain management, addressing underexplored technical, organizational, and ethical challenges. The relevance and value of this research are evidenced by its actionable insights, which can guide practitioners and policymakers in enhancing operational efficiency, innovation, and resilience in supply chains
Data Mining Analytics Approach-Based System for Crime Prediction along Kaduna–Abuja Highway, Nigeria
No full textCriminal activities in the Kaduna-Abuja Highway have been an enormous security risk to the commuters, residents, and economic growth in Northern Nigeria. Conventional law enforcement approaches are still reactive and lack data, which restricts the anticipatory and preemptive actions. This paper describes Data Mining Analytics Approach-Based System of forecasting the occurrence of crime along the Kaduna-Abuja highway, using developed models of machine learning. The analysis used three supervised learning algorithms which include Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Random Forest (RF) to represent crime patterns and predict potential types of crime, time and location. These algorithms were selected because they are widely used for classification problems, handle and have shown strong performance in similar prediction studies. Artificial data of 1,500 entries that cited five years of criminal incidences was created and prepared. Class imbalance was dealt with with the help of SMOTE (Synthetic Minority Oversampling Technique), hyperparameters optimization with the help of the GridSearchCV. The experimental outcomes showed that the KNN classifier performed best having an accuracy of 71% then SVM and Random Forest with 66.5% respectively. Analysis of the importance of features showed that the most significant predictors were month, day of the week, number of victims, year, and time of day. Study evidences that predictive analytics based on data can complement the proactive policing practices and help the security agencies to predict the high-risk times and locations along the transport corridors in Nigeria that are critical
Improving Patient Data Privacy and Authentication Protocols against AI-Powered Phishing Attacks in Telemedicine
No full textTelemedicine’s rapid expansion has improved healthcare accessibility but has also increased cybersecurity risks, particularly AI-powered phishing attacks that exploit authentication vulnerabilities. Patient data breaches are rising due to sophisticated phishing schemes targeting healthcare providers and patients. This study analyzes the impact of AI-driven phishing breaches using data from the HHS Breach Reports, Verizon DBIR, IBM Cost of a Data Breach Report, and PhishTank Open Phishing Dataset. Employing trend analysis, logistic regression, ANOVA, and machine learning classification, the findings reveal a 60% increase in patient record exposure due to AI-powered phishing since 2021, with credential theft contributing most to authentication failures (coefficient = 1.75). The study also finds that blockchain authentication reduces financial losses to 12M incurred by unprotected organizations. AI-based phishing detection achieves a recall rate of 90.5% but suffers from a 47.6% false-negative rate, indicating the need for refinement. Recommendations include implementing adaptive AI-driven threat detection, behavioral biometrics, blockchain authentication, and stronger regulatory oversight
Exploring the Use of AI-Powered Chatbots and Writing Assistants on Academic Integrity in Zambia’s Higher Learning Institutions
No full textAI-powered chatbots and writing assistants are transforming academic practices in Zambia’s higher learning institutions, raising both opportunities and challenges concerning academic integrity. These tools enhance student learning by providing instant feedback, improving writing quality, and assisting with research; however, they also pose ethical concerns related to plagiarism, authenticity, and critical thinking skills. The ease of access to AI-generated content increases the risk of academic dishonesty, as students may misuse these technologies to complete assignments without genuine effort. Hence, this study was conducted to assess the effect of AI-powered chatbots and writing assistants on academic integrity. The study adopted a mixed-methods research design, combining quantitative and qualitative approaches. The study was conducted in three higher learning institutions within Lusaka district of Zambia and sampled 345 respondents. The data collection process involved distributing the questionnaires to the selected participants and conducting individual interviews. Also, document analysis was utilized as secondary data collection tool. The data collected were analyzed using appropriate statistical methods, such as SPSS and Microsoft excel as well as research themes. The findings revealed that while these AI tools enhance students\u27 access to instant academic support, they also contribute to increased risks of academic dishonesty. Additionally, the effectiveness of institutional policies in mitigating AI-related academic misconduct remains limited due to inadequate enforcement mechanisms and lack of awareness among students and educators. On the other hand, the study also revealed limitations such as limited awareness and usage, self-reported data bias, ethical and privacy concerns, evolving AI capabilities and lack of localized AI models. Therefore, the study recommended that universities should strengthen academic integrity policies by implementing standardized AI-detection tools, conducting regular faculty and student training, and fostering a culture of academic honesty through awareness campaigns and stricter enforcement mechanisms
Enhancing Automation in QA Engineering with Advanced AI Techniques in Complex Distributed Systems
No full textAims: This study explores integrating artificial intelligence (AI) into automated quality assurance (QA) workflows for complex distributed systems.
Study Design: A multi-phase empirical approach was adopted. First, I developed a novel AI-driven test framework. Next, I deployed it in a real-world microservices environment and compared key metrics (defect detection rates, test coverage, execution time) against a conventional, manually-maintained QA suite.
Place and Duration of Study: This work was conducted at the Department of Computer Science and Engineering, «Kharkiv Aviation Institute», from January 2024 to January 2025.
Methodology: QA data (pass/fail results, defect logs, code coverage) were collected from 1,200 test cases spread across 15 microservices.
An ensemble machine learning (ML) model (Random Forest + Gradient Boosting) was trained to predict modules with high defect probability.
I integrated the AI-driven test prioritization algorithm into a Jenkins-based CI/CD pipeline.
A series of 12 iterative production releases were monitored, capturing metrics like regression test time, defect detection, concurrency handling, and QA engineer feedback. The proposed ensemble machine learning model achieved an F1-score of 0.92, reducing missed defect rates by 32% and test execution time by 45%.
Results: Test execution time reduced by 45% on average (from 110 minutes to ~60 minutes per full regression cycle).
Escaped defect rate decreased by 32%, indicating more thorough coverage of high-risk areas.
QA professionals reported a 35% increase in test efficiency and 20% fewer redundant test scripts.
Concurrency issues (e.g., thread safety, race conditions) were detected 25% earlier in the QA cycle thanks to dynamic risk-based scheduling.
Conclusion: AI-driven automation can significantly improve the speed and efficacy of QA for complex distributed systems, resulting in lower operational costs and more rapid release cycles. The proposed approach can serve as a blueprint for organizations seeking to modernize their QA pipelines with intelligent test orchestration
AI- Powered Behavioural Biometrics for Fraud Detection in Digital Banking: A Next-Generation Approach to Financial Cybersecurity
No full textThis study investigates the limitations of traditional fraud detection techniques in digital banking and explores the applicability of AI-powered behavioral biometrics as a next-generation solution for enhancing cybersecurity. Using publicly available datasets, including the PaySim Financial Transactions Dataset, Credit Card Fraud Detection Dataset, and HMOG Dataset, this research applies machine learning models such as Random Forest, Long Short-Term Memory (LSTM) networks, and Convolutional Neural Networks (CNNs). These models were evaluated using quantitative metrics including Accuracy, Precision, Recall, F1 Score, and AUC-ROC. The LSTM network demonstrated superior performance, achieving 97.9% accuracy, 95.6% precision, and 93.4% recall, outperforming other models. The results reveal that deep learning frameworks significantly enhance fraud detection efficiency, minimize false positives, and improve prediction accuracy. Furthermore, the use of publicly available datasets enhances the study’s reproducibility and transparency. Ethical considerations related to privacy, user consent, and algorithmic accountability are also discussed, highlighting the importance of responsible AI deployment in digital banking systems. This research aims to address evolving cybersecurity threats by integrating advanced deep learning models with behavioral biometrics for real-time anomaly detection. The findings demonstrate the effectiveness of AI models in accurately detecting complex fraud patterns and propose practical recommendations for integrating such systems within existing digital banking infrastructures. Recommendations include improving algorithmic transparency, establishing ethical guidelines, and investing in infrastructure upgrades to facilitate seamless implementation. This work offers a valuable foundation for future research aimed at developing robust and adaptive fraud detection systems that prioritize both efficiency and ethical compliance
AI-Driven Fault Injection Testing: Enhancing System Resilience with Automated Chaos Engineering
No full textThis paper presents a novel approach to enhancing system resilience through AI-driven fault injection testing, leveraging automated chaos engineering. As modern distributed systems grow in complexity, traditional resilience testing techniques—often limited by static fault models and insufficient adaptability—struggle to expose hidden vulnerabilities under dynamic real-world scenarios. To address these challenges, we propose an intelligent framework that integrates artificial intelligence, specifically reinforcement learning, with automated chaos tools to dynamically generate and execute context-aware fault scenarios. The system continuously learns from observed behaviors, identifies weak points, and adapts its strategies to maximize test coverage and impact. Experimental results demonstrate a 28% improvement in fault detection accuracy and a 35% reduction in system recovery time compared to conventional static methods. Furthermore, the approach generalizes effectively across various cloud-native and microservice-based architectures. This work contributes to the evolution of autonomous resilience testing, offering a scalable and proactive solution for building more robust, self-healing systems in highly dynamic environments
Ml-based Ensemble Learning Data Model for Classification Problems in Bank Marketing Prediction
No full textThis new data modelling strategy is aimed at improving predictions for telemarketing campaigns targeting potential customers for long-term deposit products at a Portuguese retail bank. The dataset includes detailed information about clients, the bank’s products, and various socio-economic factors, some of which reflect the impact of the financial crisis. Starting from an initial pool of 150 features, the model narrows this down to 21 key variables, including the target label. Our approach leverages ensemble learning and treats each feature type independently during preprocessing, followed by normalization to enhance overall predictive accuracy. To evaluate the efficiency of this technique, we compare the throughput of five widely-used classification algorithms, both individually and as part of an ensemble. The results demonstrate that integrating these techniques within an ensemble framework leads to consistently higher accuracy across all models
Plant Disease Classification of Basal Bulb Rot in Shallots Using Vision Transformer
No full textEarly detection of plant diseases is critical for sustainable agriculture and reducing crop losses. This study presents a real-time monitoring system integrating IoT and machine learning for the early detection of basal bulb rot disease in shallots. The system combines image data captured by an ESP32-CAM and soil pH data from a sensor to provide timely alerts to farmers. The images undergo preprocessing using Gaussian filtering and histogram equalization, while pH data is smoothed using a moving average filter. Features such as color, texture, shape, and pH dynamics are extracted and analyzed using a hybrid classification model comprising MobileNetV2 for image-based disease identification and Random Forest for soil pH classification, fused at the decision level. The models were optimized for edge deployment using TensorFlow Lite and field-tested under solar-powered conditions. Experimental results demonstrate a hybrid model accuracy of 93.7%, with recall and specificity of 94.1% and 92.8%, respectively. The system responds within 450 milliseconds, making it suitable for real-time applications. This solution offers a low-cost, scalable, and accurate method for precision agriculture, reducing dependence on manual inspections and enabling proactive disease management
Hybrid CNN-Haralick Framework for Foot and Mouth Disease Classification in Cattle
No full textFoot and Mouth Disease (FMD) poses a major challenge to livestock health, resulting in notable economic losses and threatening of food security. This study hereby leverages an Artificial intelligence (AI) technique by introducing a hybrid classification model that combines Convolutional Neural Network (CNN) for spatial feature extraction with texture analysis using Haralick features. Evaluated on a curated dataset of FMD-infected and healthy cattle images, the hybrid model demonstrated a notable improvement over other existing pure deep learning and CNN models, achieving an overall classification accuracy of 94%. Generally, the framework exhibited a balanced f1-score, precision and recall across all classes, addressing challenges such as overlapping patterns and class imbalance. By leveraging complementary spatial and texture-based features, the approach enhances diagnostic accuracy, offering a novel approach for FMD classification. This research underscores the value of hybrid models in advancing veterinary diagnostics and lays the groundwork for broader applications in livestock disease monitoring systems