A bayesian network based approach for software defects prediction

Since the inception of software project management, the quality of a software project is always a priority. Defects in a software development project lead to degradation of the quality which might be the cause of its failure. In order to predict the defects in a software development project we propose a Bayesian Network based approach. It is developed by using project repositories which store software metrics and defect information. In this approach we combine subjective judgements from experienced project managers and available defect rate data to produce a network and use this to forecast and hence control defect rates.