A study on software reliability prediction models using soft computing techniques

Accurate software reliability prediction can not only enable developers to improve the quality of software but also provide useful information to help them for planning valuable resources. In this paper, we examine an analytical perspective of software reliability prediction using soft computing techniques with specific focus on methods, metrics and datasets. Based on the investigated results, usage percentage of datasets of public domain and soft computing techniques has increased significantly in last ten years. However, measurements using metrics are still the most dominant methods for predicting software reliability. In practice, intelligent machine learning techniques have shown remarkable improvements for reliability prediction. Therefore, software practitioners working on software reliability prediction should continue to use public datasets and other machine learning algorithms to build better prediction models. The significant findings of our study, in conjunction with previous research, could be used as guidelines for practitioners to make predictions in more realistic operating-context.