Categorizing Code Review Comments Using Machine Learning

Arafat, Yeasir; Sumbul, Syeda; Hossain, Syeda Sumbul

doi:10.1007/978-981-16-2102-4_18

Cited by 2 publications

(1 citation statement)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Arafat et al [15] used supervised machine learning algorithms to categorize and predict the topics in code review comments from six closed-source systems and reported 63% accuracy for Support Vector Machine (SVM) method. However, their approach needs a manually labeled dataset for the initial training.…”

Section: Related Workmentioning

confidence: 99%

On potential improvements in the analysis of the evolution of themes in code review comments

Iftikhar,

Börstler,

Bin Ali

2023

2023 49th Euromicro Conference on Software Engineering and Advanced Applications (SEAA)

View full text Add to dashboard Cite

The modern code review process is considered an essential quality assurance step in software development.The code review comments generated can provide insights regarding source code quality and development practices. However, the large number of code review comments makes it challenging to identify interesting patterns manually. In a recent study, Wen et al. used traditional topic modeling to analyze the evolution of code review comments. Their approach could identify interesting patterns that may lead to improved development practices.Objective: In this study, we investigate potential improvements to Wen et al.'s state-of-the-art approach to analyze the evolution of code review comments.Method: We used 209,166 code review comments from three open-source systems to explore and empirically analyze alternative design and implementation choices and demonstrate their impact.Results: We identified the following potential improvements to the current state-of-the-art as described by Wen et al.: 1) utilize a topic modeling method that is optimized for short texts, 2) a refined approach for identifying a suitable number of topics, and 3) a more elaborate approach for analyzing topic evolution. Our results indicate that the proposed changes have quantitatively different results than the current approach. The qualitative interpretation of the topics generated with our changes indicates their usefulness.Conclusions: Our results indicate the potential usefulness of changes to state-of-the-art approaches to analyzing the evolution of code review comments, with practical implications for researchers and practitioners. However, further research is required to compare the effectiveness of both approaches.

show abstract

Section: Related Workmentioning

confidence: 99%

On potential improvements in the analysis of the evolution of themes in code review comments

Iftikhar,

Börstler,

Bin Ali

2023

2023 49th Euromicro Conference on Software Engineering and Advanced Applications (SEAA)

View full text Add to dashboard Cite

show abstract

Tutors' and Students' Views on Learning and Feedback in Problem-Based Learning

Lillo¹

2023

EUROPEAN J ED RES

View full text Add to dashboard Cite

<p style="text-align: justify;">The successful application of the problem-based learning (PBL) approach requires feedback from the tutor to the student to guide the latter's learning process. The aim of this study was to characterise the views of a group of Chilean undergraduate students of Primary Teaching and their tutors regarding PBL methodology, and the role of feedback in it. Qualitative research was carried out based on a case study that gathered the views of students and tutors through a focus group. The methods of analysis adapted to the needs of this study are those derived from discourse analysis and especially from discourse in interaction in its three dimensions: interlocution (framework of participation), thematic (topic of discussion) and enunciative (enunciative positioning). The results indicate that students' views emphasised the inter-student collaboration involved in PBL, while tutors emphasised the motivational value of the methodology. However, both recognised the important role that feedback plays in enhancing learning opportunities. Knowing the views of both tutors and learners is central to improving PBL and feedback practices.</p>

show abstract

Categorizing Code Review Comments Using Machine Learning

Cited by 2 publications

References 25 publications

On potential improvements in the analysis of the evolution of themes in code review comments

On potential improvements in the analysis of the evolution of themes in code review comments

Tutors' and Students' Views on Learning and Feedback in Problem-Based Learning

Contact Info

Product

Resources

About