SdDirM: A dynamic defect prediction model

Defect prediction and estimation techniques play an important role in software reliability engineering. This paper proposes a dynamic defect prediction model named SdDirM (System dynamic based Defect injection and removal Model) to improve the quantitative defect management process. Using SdDirM, we can simulate defect introduction and removal processes, and predict and estimate the residual defects in different phases. We describe the modeling process, the validation and the results with the empirical and real project data compared with the other well known models. These experiments show that the managers can use the model to explore and analyze the potential improvements before the practice.