Project Development Analysis of the OSS Community Using ST Mining

The OSS (Open Source Software) phenomenon is a novel, widely growing approach to develop both applications and infrastructure software recently. The fast growth of the community increases the interests in OSS related research. Accurate prediction of the project success is one of the interesting studies in OSS research. We propose to use the ST (Spatial Temporal) data mining techniques to predict the project success in the OSS community. ST mining has been studied in Euclidean distance based spatial systems like GIS, but to date has only received little attention in non-Euclidean network structured evolving system like the OSS community. In this paper, we introduce novel methods to project the evolving OSS community in a spatio-temporal data set and related ST mining algorithms to process the data set. Using ST mining techniques we propose, we are able to get the prediction of project success in the OSS community. We also present a detailed analysis and experimentally demonstrate the effectiveness and efficiency of these techniques in a real OSS community – SourceForge.net. The results show that our techniques can predict the project success and they are also useful in other non-Euclidean spatial systems. Contact: Yongqin Gao Dept. of Computer Science and Engineering University of Notre Dame Notre Dame, IN 46556 Tel: 1-574-631-7596 Fax: 1-574-631-9260 Email: [email protected]