Activity Dependency in Collaborative Networks
Pablo Loyola, Yutaka Matsuo · 人工知能学会全国大会論文集 · 2015
Open Source Software (OSS), as well as other distributed collaborative initiatives such as Wikipedia, is based on a continuous stream of voluntary contributions from developers with different incentives, skills and motivations. In this work, we focus on analyzing the behavioral patterns that emerge after a contribution is merged to the official codebase. In particular, we are interested in the activities generated by the community after they assimilate a particular contribution. This subsequent activity usually involves patches, related contributions and bug fixes, which arrive in the form of a chain reaction, especially when it is generated from a highly reputed member. We studied this phenomenon taking into account both technical and social aspects of the software development process and generated a prediction model based on the concept of cascades. Our initial experiments shows that this approach has the potential to simulate real contribution dynamics.