An efficient dual ensemble software defect prediction method with neural network

With the rapid development of technology, software projects are becoming increasingly complex, but the problem of defects is still not well solved, and the application of defective software will bring some security problems, therefore, it is necessary to identify the defective modules to ensure the quality of software. Software defect prediction (SDP) can achieve this goal and it is now an essential part of software testing. However, there is a problem of class imbalance in the defective datasets, which can easily cause the prediction models inaccuracy. Ensemble learning has been proven to be one of the best ways to address the problem of class imbalance. In this paper, we propose an efficient dual ensemble software defect prediction method with neural network (DE-SDP) to solve the class imbalance problem, thereby improving the performance of prediction model. Firstly, we combine cross-validation and seven different classifiers to build base ensemble classifiers. Then, we use stacking method and neural network model to re-ensemble the base ensemble classifiers. Finally, we evaluate the performance of proposed DE-SDP on eight public datasets, and the results demonstrate the effectiveness of the DE-SDP method.