Prediction of software quality model using gene expression programming

Abstract

There has been number of measurement techniques proposed in the literature. These metrics can be used in assessing quality of software products, thereby controlling costs and schedules. The empirical validation of object-oriented (OO) metrics is essential to ensure their practical relevance in industrial settings. In this paper, we empirically validate OO metrics given by Chidamber and Kemerer for their ability to predict software quality in terms of fault proneness. In order to analyze these metrics we use gene expression programming (GEP). Here, we explore the ability of OO metrics using defect data for open source software. Further, we develop a software quality metric and suggest ways in which software professional may use this metric for process improvement. We conclude that GEP can be used in detecting fault prone classes. We also conclude that the proposed metric may be effectively used by software managers tin predicting faulty classes in earlier phases of software development. © 2009 Springer Berlin Heidelberg.