Asymmetric Learning Based on Kernel Partial Least Squares for Software Defect Prediction

Software defect prediction is an essential part of software quality analysis and has been extensively studied in the domain of software-reliability engineering [1]–[5]. However, As pointed out by Menzies et al. [2] and Khoshgoftaar et al. [4], the class imbalance problem encountered in realworld data sets often degrades the performance of defect predictors. The software defect data set is class imbalanced when the majority of defects in a software system are located in a small percentage of the program modules. Existing approaches to solving the class imbalance problem mainly include data-level and algorithm-level methods, which are compared in [4]. Their results show that the algorithm-level method AdaBoost almost always outperforms even the best data-level methods in software defect prediction. Most recently, Qu et al. [6] proposed an asymmetric classifier APLSC, which is based on linear partial least squares, to tackle the class imbalance problem. In this paper, we develop a kernel based asymmetric learning method, called Asymmetric Kernel Partial Least Squares Classification (AKPLSC), which is able to nonlinearly extract the favorable features and retrieve the loss caused by class imbalance problem.