Mining open source software (OSS) data using association rules network
Sanjay Chawla, Bavani Arunasalam, Joseph G. Davis · 2003
Abstract. The Open Source Software(OSS) movement has attracted consider-able attention in the last few years. In this paper we report our results of min-ing data acquired from SourceForge.net, the largest open source software host-ing website. In the process we introduce Association Rules Network(ARN), a (hyper)graphical model to represent a special class of association rules. Using ARNs we discover important relationships between the attributes of successful OSS projects. We verify and validate these relationships using Factor Analysis, a classical statistical technique related to Singular Value Decomposition(SVD).