International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 04 | Apr 2025
p-ISSN: 2395-0072
www.irjet.net
A Comprehensive Review of Machine Learning Techniques in Automated Code Review Systems Ms. Anuyoksha Singh, Mr. Shrish Tiwari, Dr. Vivek Shukla MTech Scholar, Assistant Professor, Head of Department of Computer Science and Engineering, Dr. C.V. Raman University, Kota, Bilaspur, Chhattisgarh, India -------------------------------------------------------------------------***-----------------------------------------------------------------------understanding, and machine learning models for Abstract - As software development gets more
optimization and security recommendations. The system is designed to provide detailed feedback on code quality while minimizing false positives, thereby enhancing developer productivity.
complicated, strong code review and optimization methods are needed to guarantee code performance, security, and quality. Conventional manual code review techniques are uneven, time-consuming, and prone to errors. An AI-based code review and optimization system that automates code analysis, finds vulnerabilities, and makes optimization recommendations is presented in this study. By integrating ESLint for optimal analysis methods, the system's capacity to efficiently detect faults and enforce coding standards is enhanced. By streamlining the code review procedure, our suggested approach increases correctness and efficiency while lowering the amount of human labor required. A confusion matrix analysis validates the system's functionality, and rigorous testing shows that it detects code flaws with high accuracy. The system's architecture, implementation, and outcomes are highlighted in this study, demonstrating how it might revolutionize software development processes.
To create an efficient and scalable solution, we integrate Next.js and ShadCN for a dynamic and user-friendly interface, while Flask serves as the backend, managing AI interactions and processing requests. ESLint is incorporated to enforce coding best practices, ensuring consistent and maintainable code. Additionally, Cloudinary is used for file storage, allowing developers to securely manage code files and related assets.
2.REVIEW OF LITERATURE 2.1. Automatic identification of appropriate code reviewers using machine learning. Muiris, W. (2020). The work shows that a deep learning model can efficiently find suitable reviewers by learning from past pull requests. Because it was trained using features from previous pull requests, this model is better able to comprehend the context of the code changes under review.
Key Words: BERT, DISCOREV, BLEU, Artificial intelligence (AI), Convolutional Neural Networks (CNNs) , Large Language Model.
I.INTRODUCTION Software development relies heavily on code quality, which affects security, maintainability, and performance. Manual inspection is a major component of traditional code review techniques, but it is laborious, prone to human mistake, and inconsistent among reviewers. Automated solutions driven by artificial intelligence (AI) and machine learning (ML) are crucial to streamlining the review process and guaranteeing accuracy and efficiency given the growing complexity of contemporary applications.
In order to capture the subtleties of the code modifications, the model makes use of the syntactic representation of the modified code. The model can link particular code fragments to reviewers who have previously worked on comparable revisions by examining these patterns. The model's capacity to produce probabilities indicating the suitability of several reviewers for a certain pull request is one of its noteworthy results. Potential reviewers are ranked using this probabilistic method according to their prior encounters with comparable code changes.
In recent years, AI-driven tools have emerged to assist in various aspects of software engineering, including bug detection, code optimization, and security vulnerability assessment. However, existing solutions often lack flexibility, comprehensive analysis, or seamless integration into development workflows. Our proposed system addresses these limitations by leveraging Code BERT for syntax analysis, Gemini for AI-driven code
© 2025, IRJET
|
Impact Factor value: 8.315
According to the results, software development teams may find it useful to incorporate machine learning into the code review procedure. It creates opportunities for more study into improving the model and looking into other elements that can increase its efficacy and accuracy in practical situations.
|
ISO 9001:2008 Certified Journal
|
Page 1310