Comparative analysis of neural network and genetic programming for number of software faults prediction

Abstract

Software fault prediction can be more useful if, besides predicting software modules being faulty or non-faulty, number of faults can also be predicted accurately. In this paper, we present an approach to predict the number of faults in the software system. We develop fault prediction model using neural network and genetic programming and compare the effectiveness of these techniques over ten project fault datasets collected from the PROMISE data repository. The results of the prediction are evaluated using error rate, recall and completeness parameters. Our results found that for small datasets, neural network produced better results, while for large datasets genetic programming produced better results. In terms of error values, neural network outperformed genetic programming, while for recall and completeness analysis, genetic programming produced the result better than neural network.