A Method to Suggest Solutions for Software Bugs

For developers, the debugging process is a huge consumer of time and effort. They receive many bug reports that they must address by reproducing the bugs and reviewing the code for errors. Therefore, an automated system to help the developers in these tasks is very much required. This paper presents a system to find potential solutions in the Stack Overflow (SO) forum of questions and answers focusing on programming issues. It finds the similar problems, which have relevance to the reported bugs in a bug report, analyzed by the experts in SO. We combine two similarity measures to rank relevant SO posts that suggests the solutions for software bugs. Our experimental evaluations on several open-source projects show that the combination of similarity measures performs better than applying the similarity measures individually.