Agrupamento de projetos similares utilizando métricas e algoritmo BSAS

Abstract

When researchers will conduct empirical studies, one of the first tasks of an empirical research method consists of selecting which projects will be used to conduct the study. Researchers want to choose similar projects, to try to indicate a possible generalization of the results to other contexts. For each different type of study, a different set of projects may be necessary, with specific characteristics. To solve this problem there are metrics and clustering algorithms. In this work it was developed a tool that allows researchers to conduct mining software project data hosted on GitHub and select metrics pull request, user profile, log commit and project to perform the calculation with the results listed in the project, and running the BSAS algorithm to group similar projects based on their characteristics.