Data mining using machine learning techniques grabbed the interest of researchers in recent years. Software defect prediction is one of the thrust research area and data mining techniques are applied to identify the defects that are present in the datasets. In this research work an enhanced relevance vector machine (ERVM) is used for software defect prediction. From the extensive literature study it is observed that relevance vector machine classifier is comparably delivering better performance than that of support vector machine. This implication is taken for the research work. Datasets are collected from Promise software engineering repository [25] that has a collection of publicly available datasets and tools to serve researchers in building predictive software models (PSMs) and software engineering community at large. Two state of the art datasets namely PC1 and CM1 are taken for estimating the efficiency of the RVM and ERVM. MATLAB is used for implementing both RVM and ERVM. Defect prediction accuracy, sensitivity, specificity and time taken for execution are the performance metrics chosen and the results encompasses that ERVM performs better
