Collecting software defect data automatically from web site of open-source software

With the rapid development of software engineering, it is of great significance to improve the reliability of software. Only by comprehending the defects deeply can the reliability of software improve. Comprehending the defects needs a great many of software defect samples. Open-source software emerges in large numbers in recent years, and they accumulate a huge amount of information associated with software defects, which provide valuable data for the research of software defect. Therefore, an approach of extracting open-source software defect data is proposed in this research, in which the defect information and defect samples are contained. First of all, the open-source software information is obtained through the Github. Then the research of defect data extraction method based on SVM is conducted which can identify defect from obtained information. Finally, a database is established, in which the open-source software defect samples and associated information are managed. The experiment results show that the method proposed in this paper is effective and feasible.