Loading...

The Effects of Content-Based Features on Improving Code Review Automation

Sadri, Marzieh | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57087 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Fazli, Mohammad Amin
  7. Abstract:
  8. In the world of software development, Code Review is one of the most vital processes to ensure code quality and security. The textual content features in code review comments play a significant role in assessing quality and guiding the review process. This research aims to examine the importance and role of these features in identifying anti-social comments and improving code review processes. In this study, we first challenge the concept of toxicity in code review comments, which had previously been accepted as a concept in the field of code review. We focus on enhancing and automating code review processes by accurately and reliably detecting anti-social comments based on relevant features. To achieve our research goals, various methods were employed. Initially, hypothesis tests were used to challenge the comprehensiveness of the toxicity concept. Then, we used other statistical tests like analysis of variance (ANOVA) to investigate and discover relationships between anti-social features. The relationships found between anti-social features were also examined from a psychological perspective. Finally, using classical machine learning models, ensemble learning, and neural networks, we trained and evaluated model accuracy in detecting anti-social comments. The results of hypothesis tests showed that more than 20% of comments previously labeled as non-toxic are indeed anti-social, confirming the lack of comprehensiveness of the toxicity concept in code review. Additionally, the developed models in this research were able to accurately identify approximately 83.4% of anti-social comments. This research takes a significant step in rejecting the concept of toxicity and providing accurate models for detecting anti-social comments, contributing to the improvement of code review processes. The findings of this research can be utilized to enhance automated code review tools and methods, improving the efficiency and effectiveness of software development teams. This research can be seen as a fundamental step in improving culture and communication in software development environments
  9. Keywords:
  10. Machine Learning ; Code Review ; Toxic Comments Detection ; Anti-Social Comments

 Digital Object List

 Bookmark

...see more