Weighted Coupling Support (WCS): A Metric to predict the fault proneness of Object Oriented Application

Abstract

This paper presents a novel metric called Weighted Coupling Support (WCS) metric for predicting fault proneness. Our metric measures a new measurement unit called weighted coupling support for each class of the object to rank them based on their fault proneness, which is based on the well-known HITS algorithm that measures hub and authority weights in link analysis. One advantage of the proposed metric is that it can be computed in a simpler (and in many cases, programming language independent) way as compared to some of the structural metrics. We empirically studied WCS for predicting fault-proneness of classes in a large open source systems and fault prediction sensitivity compared with CBO. As the result, we found that the proposed WCS can predict fault proneness nearly with 98.7% sensitivity, which is around 42% sensible when compared to CBO.