Predicting Reassignments of Bug Reports - An Exploratory Investigation

Bug tracking systems play an important role in the software development process since they allow users to report bugs they have encountered. Unfortunately, the information provided in the bug reports may contain errors since the reporter is not always acquainted with the technical nature of the software project. It remains a manual process left to the bug triager to correct reports with erroneous information. Consequently, this results in a delay of the resolution time of a bug. In this study, we propose to use data-mining techniques to make early predictions of which particular reported bugs are assigned to the incorrect component. This way, bug triagers are assisted in their task of identifying reported bugs where the "component" field contains an incorrect value. By using open-source cases like Mozilla and Eclipse, we demonstrate the possibility of predicting which particular bug is likely to be reassigned to a different "component".