Automatic Detection of Public Development Projects in Large Open Source Ecosystems: An Exploratory Study on GitHub

Abstract

 -Hosting over 10 million of software projects, GitHub is one of the most important data sources to study behavior of developers and software projects. However, with the increase of the size of open source datasets, the potential threats to mining these datasets have also grown. As the dataset grows, it becomes gradually unrealistic for human to confirm quality of all samples. Some studies have investigated this problem and provided solutions to avoid threats in sample selection, but some of these solutions (e.g., finding development projects) require human intervention. When the amount of data to be processed increases, these semi-automatic solutions become less useful since the effort in need for human intervention is far beyond affordable. To solve this problem, we investigated the GHTorrent dataset and proposed a method to detect public development projects. The results show that our method can effectively improve the sample selection process in two ways: (1) We provide a simple model to automatically select samples (with 0.827 precision and 0.947 recall); (2) We also offer a complex model to help researchers carefully screen samples (with 63.2% less effort than manually confirming all samples, and can achieve 0.926 precision and 0.959 recall). Keywords-open source ecosystem; project sample selection; automated method; public development project;I.

An in‐depth study of the effects of methods on the dataset selection of public development projects

“…It uses greedy strategy to generate decision trees. We selected this method because it has been tested to be effective in selecting PDPs [4]. Logistic Regression (LR) .…”

“…Peril 4 still needs to be solved by researchers manually. In addition, peril 5 cannot be effectively solved by the corresponding strategy because in our previous work [4] we tested this strategy and found that this strategy cannot select PDPs with a high recall, which means that if researchers use the committer number to select project samples, they will miss many PDPs. Hence, if researchers do not want to select projects that are personal or projects that are not built for development, they have to spend considerable human effort to select samples manually.…”

“…Method 4 is a good choice in selecting a large number of PDPs [4]. Compared with the base line methods, Method 4 uses the J48 decision tree algorithm and the project description features.…”

“…blogs, projects that store the list of popular websites). That is, this method has a low precision (lower than 0.700) in selecting PDPs [4].…”

“…In our previous study [4], we proposed a machine learning‐based approach as an initial solution to this problem. Specifically, we first labelled 6369 project samples on whether these projects are PDPs or not; then, we added the basic features (e.g.…”