Analysis of Mining Techniques for Version Histories to guide software changes

Abstract

The histories of software systems hold useful information when reasoning about the systems at hand or about general laws of
software evolution. Modern software has evolved to meet the need of stakeholders, but the nature and scope of this evolution based on mining
version histories is difficult to anticipate and manage. In this paper we examine techniques which can discover interesting patterns based on
mining using association rules and training the network that can guide software developers about the changes. Mining the version histories of
software suggest and predict further likely changes and can prevent errors due to incomplete changes and provide an edge in software evolution.

Keywords: Association rule, Software Evolution, Neural Network