Abstract
Research on technical debt and community smell have drawn increasing attention in the academia of software engineering in the latest decade. Furthermore, data mining methods have been widely applied in the very domain as well. However, limited studies have contribute to the understanding of software project community using data mining methods, especially regarding the analysis of developer behaviors. Using K-means clustering, this study provides a preliminary analysis on the classification of open source software project developers based on the statistics of their behaviors related to technical debts. The results show that developers can be categorized into three different behavior groups, including, Veterans, Vulnerability Creators, and Fault Inducers/ Commoners.