Classifying Software Bug Reports Using Methods from Formal Concept Analysis

Daniel BorchmannDaniel Borchmann, Rafael PeñalozaRafael Peñaloza, Wenqian WangWenqian Wang

Download

Daniel Borchmann, Rafael Peñaloza, Wenqian Wang
Classifying Software Bug Reports Using Methods from Formal Concept Analysis
Studia Universitatis Babeş-Bolyai Informatica, 59:10-27, June 2014

Details
Bibtex

KurzfassungAbstract
We provide experience in applying methods from formal concept analysis to the problem of classifying software bug reports characterized by distinguished features. More specifically, we investigate the situation where we are given a set of already processed bug reports together with the components of the program that contained the corresponding error. The task is the following: given a new bug report with specific features, provide a list of components of the program based on the bug reports already processed that are likely to contain the error. To this end, we investigate several approaches that employ the idea of implications between features and program components. We describe these approaches in detail, and apply them to real-world data for evaluation. The best of our approaches is capable of identifying in just a fraction of a second the component causing a bug with an accuracy of over 70 percent.
Bemerkung: Note: Suplemental proceedings of the 12th International Conference on Formal Concept Analysis (ICFCA'14)
Projekt:Project: QuantLA
Forschungsgruppe:Research Group: AutomatentheorieAutomata Theory

@article{ BoPW-ICFCA14,
  author = {Daniel {Borchmann} and Rafael {Pe{\~n}aloza} and Wenqian {Wang}},
  journal = {Studia Universitatis Babe{\c{s}}-Bolyai Informatica},
  month = {June},
  note = {Suplemental proceedings of the 12th International Conference on Formal Concept Analysis (ICFCA'14)},
  pages = {10--27},
  title = {Classifying Software Bug Reports Using Methods from Formal Concept Analysis},
  volume = {59},
  year = {2014},
}