Exploring Datasets In Software Vulnerability Analysis: A Review

Computer security is a high interesting topic. Many cybercrimes have happened during the last decade as a result of exploiting software vulnerabilities. This issue has resulted in significant loss, large social concern, and increased interest in computer security. There are many methods and techniques used to find vulnerabilities in programs and applications that are written in different programming languages. The analysis method used to find vulnerabilities may be static or dynamic, and these methods depend on different datasets. In this work, a group of papers was studied on analyzing software vulnerabilities to identify the methods and techniques widely used in analysis, and it was found that most analysis methods relied on deep learning algorithms to discover vulnerabilities, which achieved good results. The vulnerabilities that were worked on were mostly buffer error, resource management error, SQLi, XSS. The SARD dataset was widely used followed by the NVD dataset, and most of the analysis methods relied on the case of static analysis and most of the programs that were worked on were in the programming languages C/C++ and Javascript.