Empirical Analysis of Artificial Immune System Algorithms for Aging Related Bug Prediction

The complex nature of human immunology algorithms has motivated the research community to explore their practical applications in various other fields. As a result, Artificial Immune Systems (AISs) is one such class of algorithms that has found its way into software quality predictive modeling. In this paper, we evaluate AIS algorithms for developing Aging-Related Bug (ARB) prediction models. Software Aging, the gradual degradation and resource exhaustion in software systems, is said to be caused by ARBs, which may or may not be identified during software testing. Therefore, predicting ARBs before software release can help software managers in reducing their impact. This paper presents an empirical study that statistically analyzes the effectiveness of AIS classifiers for ARB prediction on five open-source software datasets. In order to account for the imbalanced nature of the investigated datasets, we used resampling and cost-sensitive classifiers. The results of the study indicate the effectiveness of AIS algorithms for developing ARB prediction models.