A Commit Change-based Weighted Complex Network Approach to Identify Potential Fault Prone Classes
Chun Yong Chong, Sai Peck Lee · 2018
Over the past few years, attention has been focused on utilizing complex network analysis to gain a high-level abstraction view of software systems. While many studies have been proposed to use interactions between software components at the variable, method, class, package, or combination of multiple levels, limited studies investigated how software change history and evolution pattern can be used as a basis to model software-based weighted complex network. This paper attempts to fill in the gap by proposing an approach to model a commit change-based weighted complex network based on historical software change and evolution data captured from GitHub repositories with the aim to identify potential fault prone classes. Experiments were carried out using three open-source software to validate the proposed approach. Using the well-known change burst metric as a benchmark, the proposed method achieved average precision of 0.77 and recall of 0.8 on all the three test subjects.