A Novel Recommendation System for Code Validation and Optimal Refactoring (Nrscvor)

Authors

K. Pranith Chowdary Google Scholar

Abstract

Software has become an integral part of our daily lives, impacting almost every aspect of modern society. However, with this increased reliance on software, a need to ensure its reliability, security, and ethical use also comes. In our work, we have highlighted the importance of writing clear and simple code to read, modify, and extend, and follows best practices and coding standards. Hence, we designed and developed a novel Recommendation System (RS) that integrates AI-based techniques to assess the overall quality of the code and recommend precise and comprehensible code for developers. The effectiveness of our proposed model is measured in terms of evaluation metrics like accuracy, recall, precision and F1-Score, along with other metrics such as Mean Average Precision for 100 samples (mAP@100), mAP@50, and Bilingual Evaluation Understudy 4 (BLEU-4) and compared with the existing systems like codeBERT, GraphCodeBERT, DeepCom etc. According to the experimental results, the proposed system performs better than the existing systems.

Keywords: Code smell detector, Recommender system, code review, similarity score

This paper was under review in the Journal titled "Information and Software Technology" by ScienceDiect under Elsevier.

Journal Info

• Subject areas : Information Systems, Software, Human-Computer Interaction
• Impact : Cite Score : 9.2 | Impact Factor : 3.9
• Publishing timeline : 233 Days (Submission to Acceptance) | 8 days (Acceptance to Publication)
• Abstracting and indexing :

  1. Web of Science

  2. Ergonomics Abstracts

  3. INSPEC - IT-Digest

  4. Science Citation Index Expanded

  5. ACM Guide to Computing Literature

  6. Applied Science and Technology Index

  7. Computer Literature Index

  8. CompuScience

  9. Current Contents

  10. Deadline Newsletter

  11. Engineering Index

  12. Research Alert

  13. Scopus

      ---

Link to the Preprint of this paper

Sugestted Citation : Karumanchi, Pranith Chowdary and Dasari, Yakobu and Kolli, Venkata Krishna Kishore and Mothadakala, Ramya, A Novel Recommendation System for Code Validation and Optimal Refactoring (Nrscvor). Available at SSRN: https://ssrn.com/abstract=4474677 or http://dx.doi.org/10.2139/ssrn.4474677