ClassifyHub: An Algorithm to Classify GitHub Repositories

Marcus Soll, Malte Vosgerau

Abstract: The classification of repositories found on GitHub can be considered as a hard task. However, the solution of this task could be helpful for a lot of different applications (e.g. Recommender Systems). In this paper we present ClassifyHub, an algorithm based on Ensemble Learning developed for the InformatiCup 2017 competition, which is able to tackle this classification problem with high precision and recall. In addition we provide a data set of classified repositories for further research.