Hybrid depth defect prediction method based on code snippet analysis

Abstract

The invention relates to a hybrid depth defect prediction method based on code snippet analysis, and belongs to the technical field of computer software defect prediction. According to the method, firstly, based on a program slicing method of defect library key points, an open source software code unit set containing defects is vectorized, and features are expressed as a vector form which can be processed by a deep learning model; then, based on a defect prediction method of hybrid deep learning, the classification and prediction capabilities of a hybrid deep model are improved, and a defect prediction classifier is obtained through training; and finally, defect prediction is performed on the open source software based on the trained defect prediction classifier, and target code snippets are outputted in a classified manner. According to the method, the pre-designed defect library key points are taken as the entry point of program slicing, the code snippets containing defect characteristics are extracted from the open source codes and are vectorized and expressed; and the hybrid model is obtained based on the multiple deep learning methods, so that the data processing capability and the automatic learning capability of the model can be effectively improved.

Description

technical field [0001] The invention relates to an open source software hybrid deep defect prediction technology based on code segment analysis, and belongs to the technical field of computer software defect prediction. Background technique [0002] With the development of computer technology, especially the rise of Internet technology and related enterprises, open source software has become the mainstream in various aspects such as research and development framework, operating system, compilation tool chain, database, and WEB server. In the informatization construction of major enterprises and institutions, the application scope of open source software has gradually expanded. Open source components that have been widely used include: R&D frameworks (such as Spring, Struts, etc.), operating systems (such as ubuntu), databases (such as Mysql), middleware (such as Tomcat), functional components (such as Zookeeper, Solr), etc. Open source software has the characteristics of op...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the technical problems of uneven code quality and potential safety hazards in open source software, and propose a method for predicting mixed depth defects of open source software based on code fragment analysis to predict whether the input code unit set contains defects , to ensure the security of using open source software

Images