Feature Selection using Genetic Algorithm for Software Fault Prediction

The field of software engineering has seen an increase in research aimed at predicting software defects. The importance of non-functional defects in this study can effectively predict software defects, thus simplifying software testing, reducing costs, and improving the overall software testing process. However, the presence of noise in the data set often causes software bug prediction performance to decrease. In this work, a Genetic Algorithm (GA) is used to improve the accuracy of software fault prediction. It is used as a feature selection (FS) algorithm. The technology was evaluated using NASA’s Metrics Data Repository dataset. The results show that the prediction performance of most products is greatly improved when the proposed method is used.