Software Defect Prediction based on Adaboost algorithm under Imbalance Distribution

摘要

Software defects will lead to software running error and system crashes. Many methods were proposed to solve this problem. However, the imbalance distribution of software defects leads to the major bias and accuracy loss for most software defect prediction methods. In this paper, we propose an application which combine Adaptive Boosting (AdaBoost) and Back-propagation Neural Network (BPNN) algorithm to train software defect prediction model. BPNN was utilized as a weak leaner in AdaBoost and tweaked in favor of instances misclassified. The experiments show that the proposed method in the paper significantly improves the performance than the previous models, which is effective to deal with the imbalance software defect data.