Research on software defect prediction based on data mining

Abstract

As the development of computer technology, software system becomes more and more complicated. Because of human's ability limit, there must be a lot of defects generated in the software development life cycle. This paper reviewed the state of art in the field of software defect management and prediction, and presented data mining technology briefly. Finally, proposed an ideal software defect management and prediction system, researched and analyzed several software defect prediction methods based on data mining techniques and specific models (Bayesian Network and PRM). With this system, we can efficiently draw up some prevention and solution scheme to guide the development of new software.