On combining IR methods to improve bug localization

Document Type

Conference Proceeding

Publication Date

7-13-2020

Abstract

Information Retrieval (IR) methods have been recently employedto provide automatic support for bug localization tasks. However,for an IR-based bug localization tool to be useful, it has to achieveadequate retrieval accuracy. Lower precision and recall can leavedevelopers with large amounts of incorrect information to wadethrough. To address this issue, in this paper, we systematically investigate the impact of combining various IR methods on the retrievalaccuracy of bug localization engines. The main assumption is thatdifferent IR methods, targeting different dimensions of similaritybetween artifacts, can be used to enhance the confidence in eachothers' results. Five benchmark systems from different applicationdomains are used to conduct our analysis. The results show that a)near-optimal global configurations can be determined for differentcombinations of IR methods, b) optimized IR-hybrids can significantly outperform individual methods as well as other unoptimizedmethods, and c) hybrid methods achieve their best performancewhen utilizing information-theoretic IR methods. Our findings canbe used to enhance the practicality of IR-based bug localizationtools and minimize the cognitive overload developers often facewhen locating bugs.

Publication Source (Journal or Book title)

IEEE International Conference on Program Comprehension

First Page

252

Last Page

262

This document is currently not available here.

Share

COinS