| The past decade has seen a surge of interest in the network representation of a wide range of real world systems, from social relationships among individuals, to interactions of proteins in biological systems. The interdependence of components, including both developer and code base, in large software systems as prototypes of complex networks have attracted much research attention. This dissertation presents the application of complex network theory in Open Source Software(OSS) projects, which helps identify refactoring event and leads to a better understanding of the structure OSS projects and their developers. Furthermore, how a network breaks up into sub-networks or communities is of wide interest. Here, we show that vertices connected to many other vertices across a network can disturb the community structures of otherwise ordered networks, introducing noise. We investigate strategies to identify and remove noisy vertices ("violators"), and develop a quantitative approach to identify when the largest enhancement to a modularity measure is achieved. We show that removing nodes thus identified reduces noise in detected community structures for a range of different types of real networks in software systems and in biological systems. Co-evolution of social and technical networks in software systems is investigated as well, which can be extended to other two-layered or bipartite networks. We also present applications of different network metrics in failure prediction for software systems. By applying both unsupervised clustering and supervised classification algorithms, we show that network metrics provide additional predictive power compared with traditional software metrics in failure prediction. |
