A Software Defect Prediction Approach based on Machine Learning

In recent years, there has been an increasing interest in using machine learning methods to predict software defects. In practical engineering applications, the number of defective software samples is far less than that of non-defective software samples, resulting in unbalanced distribution of samples. In addition, the prediction performance of a single classifier is limited due to the diversity of distribution of defect data. Therefore, a series of methods should be adopted to solve these problems to improve the accuracy and reliability of software defect prediction. Here, the combined sampling method is used to deal with the class imbalance problem, and the minority class defect samples are synthesized. Through improved Bagging, different base classifiers are selected to increase the diversity of classifiers in defect prediction, and the final prediction result is obtained by weighted integration of multiple classifiers according to the Accuracy value of each classifier for the test set classification. Through the above method, a better classification effect is obtained.