Recommending GitHub Projects by Leveraging Developers' Social Networks and Genetic Algorithm

Abstract

In the uncommon and challenging year 2020, face-to-face collaboration is difficult.In particular, in the domain of software, online collaboration becomes more demanding than ever. As the leading platform of open-source software, GitHub, without a doubt, plays a vital role in hosting over 190 million of projects and fostering collaboration among developers. Developers in GitHub continuously seek new opportunities to contribute to new projects. So far, several researchers have used users' historical activities, textual descriptions of projects or starred items to analyze or infer the interests or programming expertise of developers to offer possible project recommendations. While some research utilized social connections to obtain developers' social importance, none of them have implemented such aspects for generating project recommendations. In this research, we use the latest GHTorrent dataset to construct and propose a GitHub project recommendation system by leveraging Parallel Genetic Algorithm and developers' social networks. To the best of our knowledge, this is the first application of Genetic Algorithm in the GitHub project recommendation area. iii I, Lance PoKai Wang, would like to express my sincere gratitude to my fantastic supervisor, Professor Olga Baysal, for her continuous guidance, advice, and friendly discussions throughout my thesis. Her precise insights and constant effort provided me valuable feedback and support, which made this work successful. I also thank Professor Frank Dehne for his advice and guidance on constructing parallel frameworks and also suggesting to work weights-distributed structure. I am also grateful to Adam Spriggs. He has contributed to this work by implementing the baseline RepoPal, and his work has speed up the model evaluation.I also thank my laboratory members, especially Norbert Eke and Khadija Osman, for providing insightful feedback for data collection and preprocessing. I also wish to express my deepest appreciation to all the thesis committee members for their valuable time, insightful comments and encouragement.Last but not least, I would like to thank my mother, LiChin, my brother, PoJun for their endless love and support in my entire life. v v Table of Contents vi List of Tables x List of Figures xi Nomenclature xiii List of Tables