Intelligent Radar Software Defect Prediction Approach and Its Application

Abstract

Radar software defects are not used and applied effectively and sufficiently in the testing process. As a result, defects often occur repeatedly, which causes safety hazards in software operation. To resolve this problem, this paper proposed a novel intelligent defect prediction approach for radar software by using Naïve Bayesian to classify defect data and predict defects according to radar software requirements. We apply the proposed approach on the typical radar software. The experiment results show that the defect prediction precision rate of the proposed defect prediction approach is 75%, and the prediction recall rate is 70%, approx. The experiment results are better compared to the defect prediction methods without Naïve Bayesian and defect classification. Therefore, the proposed defect prediction approach can be applied on radar software effectively and applicably to improve the effectiveness of radar software testing and provide the positive feedback to the radar software design process significantly.