Use of Bug Databases for Effort Estimation

Rahul Premraj (Univ. of Saarbrucken)

A project manager is often required to schedule the release of a product that its nearing the end of its development stage. During this final development period, most efforts are concentrated upon fixing identified bugs and code refactoring. While the latter is of lesser priority, the former is critical for the success of the software. Hence, to be able to schedule the release, the manager must have the ability to predict how long would it take to fix all reported bugs in the product. We present a novel technique for making such estimations by leveraging the data stored in bug databases that are used by many open and closed source software for tracking bugs in their products. Our technique was used to predict time estimates for fixing new bugs. Initial results show that our performance was at par with experts' estimates.