Software Defect Regions Identification Based on Behavior Mining

Abstract

Abstract. Random testing is a widely used testing technique, and adaptive random testing has enhanced the performance of random testing recently. The research results demonstrate that the improvement depends on the characters of the software defect region. A method that depicts the distribution of software defect region in the input region based on testing is proposed in this paper, and the method that describes the characteristics of the software defect region. The experimental results show that testing constraint reveals the mechanism of software defect trigging and propagation.