Automatic bug assignment has been a topic of interest for researchers in recent years. Bug reports play a crucial role in helping engineers locate and fix bugs in software systems. However, the presence of noise in textual bug reports can negatively affect the effectiveness of automatic bug assignment systems.
A research team led by Zexuan Li recently published their findings in Frontiers of Computer Science. The team sought to understand the impact of textual features and nominal features on bug assignment approaches. They focused on the effectiveness of TextCNN, a deep learning-based NLP technique, in handling textual features.
Contrary to expectations, the study revealed that textual features did not outperform nominal features in bug assignment approaches. The team discovered that nominal features, which reflect developer preferences, were more influential in improving bug assignment accuracy. By utilizing the wrapper method and bidirectional strategy, the team identified key features that significantly contributed to bug assignment.
Experimental Results
The experiments conducted by the research team involved training models with fixed classifiers on varying groups of features. The results showed that nominal features, when used strategically, can enhance bug assignment accuracy by 11-25% under popular classifiers such as Decision Tree and SVM. This highlights the importance of considering nominal features in bug assignment approaches.
Future Implications
The research suggests that future work in bug assignment approaches should focus on incorporating source files to establish a knowledge graph. This graph can facilitate a deeper understanding of the relationship between influential nominal features and descriptive words. By leveraging this relationship, bug assignment systems can be further optimized for improved performance.
Overall, the study underscores the significance of nominal features in bug assignment approaches and challenges the assumption that textual features alone are sufficient for effective bug assignment. The findings open up new avenues for research in the field of automatic bug assignment, emphasizing the importance of considering developer preferences and utilizing advanced techniques to enhance bug assignment accuracy.
Leave a Reply