BAHA: A Novel Approach to Automatic Bug Report Assignment with Topic Modeling and Heterogeneous Network Analysis
Wen ZhangSchool of Economics and ManagementBeijing University of Chemical TechnologyBeijing100019ChinaSong WangDepartment of Electrical and Computer Engineering atUniversity of WaterlooOntarioN2L 3G1CanadaQing WangLaboratory of Internet TechnologiesInstitute of SoftwareChinese Academy of SciencesBeijing100190China
ABI
Abstract
We propose an approach called Bug report assignment with topic modeling and heterogeneous network analysis (BAHA) to automatically assign bug reports to developers. Existing studies adopt social network analysis to characterize the collaboration of developers. The networks used in these studies are all homogenous. In real practice of bug resolution, different developers collaborate on different bug reports that makes the homogenous network unable to capture this information. We use heterogeneous network to describe the relations between reporters, bug reports and developers to characterize developers' collaboration. Experiments on Eclipse JDT project show that BAHA outperforms the state of art methods on automatic bug report assignment.
Topics
Identifiers
Citations and references
Cited by 00 references
Metrics — AkademScholar · Coming soon