Context: Cross-project defect prediction (CPDP) research has been popular. One of the techniques for CPDP is a relevancy filter which utilizes clustering algorithms to select a useful subset of the cross-project data. Their performance heavily relies on the quality of clustering, and using an advanced clustering algorithm instead of simple ones used in the past studies can contribute to the performance improvement. Objective: To propose and examine a new relevancy filter method using an advanced clustering method DBSCAN (Density-Based Spatial Clustering). Method: We conducted an experiment that examined the predictive performance of the proposed method. The experiments compared three relevancy filter methods, namely, Burak-filter, Peters-filter, and the proposed method with 56 project data and four prediction models. Results: The predictive performance measures supported the proposed method. It was better than Burak-filter and Peters-filter in terms of AUC and g-measure. Conclusion: The proposed method achieved better prediction than the conventional methods. The results suggested that exploring advanced clustering algorithms could contribute to cross-project defect prediction.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.