Problems with SZZ and Features: An empirical study of the state of practice of defect prediction data collection
This work addresses data quality issues in defect prediction for software engineering, highlighting severe validity threats from inaccurate labels, but it is incremental as it builds on known problems with SZZ and feature sets.
The study analyzed the SZZ algorithm for labeling bug-fixing commits in defect prediction data, finding that only half of its identified commits are actually bug-fixing, and using it with a six-month time frame leads to one incorrect label per correct label while missing two defective files. It also found that using more features beyond static code metrics does not significantly impact results.
Context: The SZZ algorithm is the de facto standard for labeling bug fixing commits and finding inducing changes for defect prediction data. Recent research uncovered potential problems in different parts of the SZZ algorithm. Most defect prediction data sets provide only static code metrics as features, while research indicates that other features are also important. Objective: We provide an empirical analysis of the defect labels created with the SZZ algorithm and the impact of commonly used features on results. Method: We used a combination of manual validation and adopted or improved heuristics for the collection of defect data. We conducted an empirical study on 398 releases of 38 Apache projects. Results: We found that only half of the bug fixing commits determined by SZZ are actually bug fixing. If a six-month time frame is used in combination with SZZ to determine which bugs affect a release, one file is incorrectly labeled as defective for every file that is correctly labeled as defective. In addition, two defective files are missed. We also explored the impact of the relatively small set of features that are available in most defect prediction data sets, as there are multiple publications that indicate that, e.g., churn related features are important for defect prediction. We found that the difference of using more features is not significant. Conclusion: Problems with inaccurate defect labels are a severe threat to the validity of the state of the art of defect prediction. Small feature sets seem to be a less severe threat.