82 Automatic Bug Assignment Using History of Packages Available to Purchase

In the process of software debugging, a valid reported bug, after validation checking, should be assigned to a developer to fix. This paper presents an automatic approach to ease the bug assignment stage of the bug triage process. Our approach comprises two stages. In the first stage, the package to which the changed file to fix the bug belongs is predicted using the machine learning algorithm; and in the second stage, the most appropriate developer is recommended using the package activity histories of the predicted package. The aim of this paper is recommending the most appropriate developer to fix a new bug. We reach recall levels of 31% on the JDT component of Eclipse project.