UNDERSTANDING DEVELOPER RELATIONS IN FOSS ECOLOGY USING SNA PARALLELIZATION TECHNIQUE

Abstract: Social network analysis aims at understanding the organization of a social network at various levels. This paper involves analysis of BigData related to social interaction among developers in Free and Open Source Software ecosystem. The analysis involves spotting influence, predicting future links and Clustering entities. The analysis was first made sequentially. However, as the size of the data kept increasing sequential computing was inefficient. In this paper, the analysis has been taken forward in parallel implementation on multi-node Hadoop cluster to improve computation time. Effective performance benefits have been achieved by considering the terms that affect the performance.

Keywords: Influence, Link Prediction, Proximity measure, Spanning Tree, Clique, Purity.