Journal of Computer Science Engineering and Software Testing

An AI-Driven Predictive Framework for Intelligent Test Prioritisation in CI/CD Environments

Wed, 14 Oct 2026 00:00:00 +0000

Software quality has become essential for ensuring modern applications perform reliably amid increasing complexity and feature richness, where robust testing is critical for security, performance, and user-friendliness. The discussed framework leverages artificial intelligence to predict and prioritize tests within CI/CD pipelines, integrating core engineering principles with advanced testing strategies to streamline processes and elevate overall application quality. Automation tools enhance traditional testing approaches, such as unit testing for individual components, integration testing for interactions, system testing for end-to-end functionality, and user acceptance testing for user alignment, by reducing manual effort. At its core, the model's machine learning component analyzes historical defect data to identify high-risk areas, enabling smarter prioritization that accelerates defect detection during development cycles. To validate the framework, a dedicated web application was developed with curated test scenarios, executed test cases, and comprehensive performance and reliability assessments. The results demonstrated significant time savings in testing efforts, higher defect detection rates, improved maintainability, and seamless compatibility with continuous delivery practices. By embedding AI-driven predictions into the software development lifecycle, organizations can achieve more scalable, high-quality software while sustaining CI/CD efficiency. Although certain implementation details may appear intricate or subject to varied interpretations, this approach ultimately fosters the creation of dependable, resilient systems.

AI-Driven Smart Agriculture Platform for Integrated Crop Health Monitoring and Yield Optimization: A Systematic Literature Review

Anitha L., Rakshith Prabhu, Srihari M. R., Srinivas Gowda C. N., Suhas C. — Mon, 14 Sep 2026 00:00:00 +0000

Agriculture is undergoing rapid transformation through the integration of Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), and Internet of Things (IoT) technologies. These advancements enable intelligent crop disease detection, crop yield prediction, and precision farming practices that improve agricultural productivity while minimizing resource consumption. Recent studies have demonstrated significant improvements in disease classification accuracy using Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), and multimodal learning techniques. Similarly, advanced regression and ensemble learning algorithms have enhanced crop yield prediction by utilizing climatic, soil, and environmental parameters. This systematic literature review analyzes recent research published between 2021 and 2026 on AI-driven smart agriculture systems. The review compares various machine learning algorithms, datasets, evaluation metrics, and their advantages and limitations reported in the literature. Furthermore, the study identifies current research gaps and discusses future research opportunities involving Explainable Artificial Intelligence (XAI), multimodal learning, federated learning, and real-time agricultural decision support systems. The findings indicate that integrating multiple AI techniques into a unified smart agriculture platform can significantly improve prediction accuracy, transparency, and practical usability for modern precision farming.

Secure Data Warehousing through Access Control: Critical Challenges, Limitations, and Future Perspectives

Aamir Khan, Arpit Tyagi, Sanjiv Kumar, Gulshan Sachan, Akanksha, Aanchal — Sat, 17 Oct 2026 00:00:00 +0000

Critical review of data warehouses that store large volumes of sensitive, heterogeneous, and business-critical information, making effective access control essential for maintaining confidentiality, integrity, and availability. Traditional access control mechanisms, including Discretionary Access Control, Mandatory Access Control, Role-Based Access Control, and Attribute-Based Access Control, have been widely adopted to regulate access to organizational data. The increasing scale, complexity, distribution, and dynamic nature of modern data warehouse environments have exposed several limitations in conventional approaches. This paper presents a critical analysis of traditional access control mechanisms with particular emphasis on their security limitations, operational challenges, and emerging requirements. The study examines their effectiveness in addressing dynamic user behavior, fine-grained authorization, evolving access patterns, insider threats, privilege escalation, scalability, and context-aware access requirements. The findings indicate that although traditional mechanisms provide structured and policy-driven authorization, they often rely on predefined rules and static configurations, limiting their ability to respond effectively to continuously changing access behaviors and emerging security threats. The analysis further highlights challenges related to policy management complexity, adaptability, real-time decision-making, and behavioral anomaly detection. Based on these findings, the paper identifies emerging requirements for more adaptive, intelligent, context-aware, and behavior-sensitive access control mechanisms.

Automated Detection and Prioritization of Civic Infrastructure Complaints Using a Cascaded YOLOv8–LLM Pipeline with Geospatial Priority Scoring on Serverless AWS

Fri, 25 Sep 2026 00:00:00 +0000

The systems used for lodging complaints by the municipalities in most Indian cities involve manual indenting, which leads to delays, inaccurate information, and wrong routing of complaints to departments. This paper presents a serverless web application that automates all the stages in the civic complaint life cycle, which prolongs the whole process. In this system, the complaint is submitted through images only, and the citizens do not need to define or classify the complaint. The important aspect of this system is that it works based on the seven-phase AI pipeline, which incorporates a specially trained YOLOv8 detector to determine 4 most common types of complaints: potholes, garbage dumps, waterlogging, and faulty streetlights. In case of low confidence prediction can always fall back on Amazon Nova Lite, a multimodal large language model that can be accessed using Amazon Bedrock. Furthermore, the pipeline also performs the assessment of the severity of the complaint, automatic routing to a department, generation of complaint descriptions, and priority scoring. The backend part of the application is executed by means of AWS Lambda, S3, API Gateway, and DynamoDB. The mentioned YOLOv8 has an mAP@50 indicator equal to 0.75 for all 4 classes.