Predicting the priority of bug reports using classification algorithms

Abstract

Severity of the bug reports are assigned by the user, developer and tester of the software, but priority is assigned by the developers. The developer assigns different priorities such as P1, P2, P3, P4, and P5 where P1 is higher priority and P5 is lower priority. If a bug report contains the priority level P1, then it will be given higher priority for fixing. In order to prioritize which bug to be fixed first based on the priority, priority information is required. Even though, the priority is assigned by developer, sometimes it may incorrect, because of busy schedule or inexperienced developer. That time, developer can use this recommendation system for more accurate priority assignment and also time may be saved. In this work, predicting the priority of bug report is presented using different classification algorithms such as Naïve bayes, Simple Logistic and Random Tree. Among the three classifiers Simple Logistics gives better result over other two classifiers.