Analyzing and Modeling Open Source Software Bug Report Data

We analyzed the major differences between closed and open source software development from a software reliability perspective. We examined real world bug report data from six open source software projects, as well as the relationship between the release cycles and the stability characteristics of the bug report data of open source software projects. We then modeled the bug report data using nonparametric techniques. The experimental results suggest that generalized additive models and exponential smoothing approaches are suitable for the estimation of software reliability at least for some of the open source projects.