Software defect prediction method, device, storage medium and electronic equipment

Abstract

The invention discloses a software defect prediction method, a software defect prediction device, a storage medium and electronic equipment. The method comprises: selecting a predetermined number of sample data in a first preset original data set according to a first preset selection rule to obtain a first prototype data set; calculating a first data set of dissimilarity between the first preset original data set and the first prototype data set according to the first preset distance algorithm; and inputting the data in the first data set to a preset software defect prediction model to obtainthe software defect prediction result corresponding to the first preset original data set, wherein the preset software defect prediction model is a model constructed according to the preset dissimilarity. By using the invention, it is possible to determine whether the software has defects according to the dissimilarity, and fundamentally improve the prediction performance with relatively high accuracy and good user experience, thereby solving the problems in the prior art.

Description

technical field [0001] The invention relates to the field of data processing, in particular to a software defect prediction method, device, storage medium and electronic equipment. Background technique [0002] The number of defective samples in software defect datasets is often much smaller than that of non-defective samples, so software defect prediction can be regarded as a class imbalance learning problem. In the learning process of class imbalance learning, the misclassification costs of different classes are also different, and the misclassification cost of the minority class (defective) is much higher than the misclassification cost of the majority class (non-defective). For the cost of misclassification, the prediction algorithm pays more attention to improving the prediction accuracy of defective minority samples. In fact, traditional classification algorithms are usually based on the premise of balanced class distribution and equal misclassification costs, with th...

AI Technical Summary

Problems solved by technology

[0007] The present invention provides a software defect prediction method, device, storage medium and electronic equipment to solve the following problems in the prior art: the existing class imbalance learning method focuses on how to adjust the class distribution or improve the algorithm, and cannot fundamentally Improve the prediction performance of this type of problem, the prediction accuracy is low, and the user experience is poor

Images