ABSTRACT
Static complexity metrics are often used to measure the functional complexity of software modules. Classifying software modules, based on their static complexity measures, into different fault-prone categories is an important problem in software engineering. This research investigates the applicability of neural network classifiers for identifying fault-prone software modules using a data set from a commercial software system. A preliminary empirical comparison is performed between a minimum distance based Gaussian classifier, a perceptron classifier and a multilayer feed-forward network constructed using a modified Cascade-Correlation algorithm. Our preliminary results suggest that multilayer feed-forward networks can be used as a tool to identify fault-prone software modules early during the development cycle.
