A Machine Learning Approach for Classifying the Default Bug Severity Level.
Saved in:
| Title: | A Machine Learning Approach for Classifying the Default Bug Severity Level. |
|---|---|
| Authors: | Aburakhia, Abdalrahman1, Alshayeb, Mohammad1,2 alshayeb@kfupm.edu.sa |
| Source: | Arabian Journal for Science & Engineering (Springer Science & Business Media B.V. ). Sep2024, Vol. 49 Issue 9, p13131-13148. 18p. |
| Subjects: | Defect tracking (Computer software development), Default (Finance), Support vector machines, Feature extraction, Sentiment analysis, Software maintenance |
| Abstract: | Bug reports (BRs) play a major role in the software maintenance process; they alert developers about the bugs discovered by the end-users. Software applications utilize bug tracking systems (BTS) to manage submitted bug reports. Recent studies showed that the majority of BRs in BTS belong to the default severity category, which does not represent their actual severity. In this paper, we propose an approach that can automatically classify default bug reports into severe or non-severe categories. We curated a dataset based on the history of bug report logs. After that, we used the Support Vector Machine algorithm and Term Frequency-Inverse Document Frequency feature extraction method to classify default bug reports into severe or non-severe categories. The results show that building customized models for default severity bug reports provides better and more reliable results than training one model for all severity. Overall, the proposed Log model outperformed the three models (approaches) from the literature; it achieved an improvement of up to ~ 4% f-measure compared to others, and in some projects, it achieved an improvement of 11.2% f-measure. Moreover, we investigated the impact of sentiment analysis on default bug severity prediction; the results show no noticeable influence. [ABSTRACT FROM AUTHOR] |
| Copyright of Arabian Journal for Science & Engineering (Springer Science & Business Media B.V. ) is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Bug reports (BRs) play a major role in the software maintenance process; they alert developers about the bugs discovered by the end-users. Software applications utilize bug tracking systems (BTS) to manage submitted bug reports. Recent studies showed that the majority of BRs in BTS belong to the default severity category, which does not represent their actual severity. In this paper, we propose an approach that can automatically classify default bug reports into severe or non-severe categories. We curated a dataset based on the history of bug report logs. After that, we used the Support Vector Machine algorithm and Term Frequency-Inverse Document Frequency feature extraction method to classify default bug reports into severe or non-severe categories. The results show that building customized models for default severity bug reports provides better and more reliable results than training one model for all severity. Overall, the proposed Log model outperformed the three models (approaches) from the literature; it achieved an improvement of up to ~ 4% f-measure compared to others, and in some projects, it achieved an improvement of 11.2% f-measure. Moreover, we investigated the impact of sentiment analysis on default bug severity prediction; the results show no noticeable influence. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 2193567X |
| DOI: | 10.1007/s13369-024-09081-8 |
