Fuzzy Integral Based on Mutual Information for Software Defect Prediction

Software defect prediction is an important research direction in the field of software reliability. In the process of software defect prediction, the interaction among the various attributes will affect classification performance of classifier. In order to fully reflect the importance of attributes and the interaction among the attributes in the process of classification, the paper proposes a fuzzy integral method based on the mutual information (MIFI) to predict the software defect. The algorithm uses the mutual information between the various attributes to determined the fuzzy measure set function, which reflect the attributes information and interaction information among attributes. The experimental results show that MIFI algorithm proposed in this paper can achieve better prediction effect than other three methods.