Assessing the Fault Proneness Degree (DFP) by Estimating the Impact of Change Request Artifacts Correlation

TL;DR

This paper introduces a novel approach called DFP-CRC that assesses fault-proneness of change requests by analyzing the impact of artifact correlations, enhancing fault prediction accuracy in software maintenance.

Contribution

It extends previous models by incorporating change request artifact correlations into fault-proneness estimation, providing a more comprehensive assessment method.

Findings

Effective in predicting fault-proneness using correlation analysis

Improves fault prediction accuracy over previous models

Validated on real-world maintenance project data

Abstract

Exploring the impact of change requests applied to a software maintenance project helps to assess the fault-proneness of the change request to be handled further, which is perhaps a bug fix or even a new feature demand. In practice, the major development community stores change requests and related data using bug tracking systems such as Bugzilla. These data, together with the data stored in a versioning system, such as Concurrent Versioning Systems, are a valuable source of information to create descriptions and also can perform useful analyzes. In our earlier work, we proposed a novel statistical bipartite weighted graph-based approach to assessing the degree of fault-proneness of the change request and Change Request artifacts. With the motivation gained from this model, here we propose a novel strategy that estimates the degree of fault-proneness of a change request by assessing the impact of a change request artifact towards fault-proneness that considers the correlation between change requests artifact as another factor, which is in addition to our earlier strategy. The proposed model can be titled as Assessing the Fault Proneness Degree of Change Request Artifacts by estimating the impact of change requests correlation (DFP-CRC). As stated in our earlier model, the method DFP-CRC also makes use of information retrieval methods to identify the change request artifacts of the devised change request. And further evaluates the degree of fault-proneness of the Change Requests by estimating the correlation between change requests. The proposed method is evaluated by applying on concurrent versioning and Change request logs of the production level maintenance project.