Fuzzy Logic Based Model to Predict Per Phase Software Defect

Abstract

Software reliability is expressed as the probability of software to function properly under specified condition for a specified time period. A basic method to evaluate the software reliability is to check the presence of defects in the software. The presence of defect can be calculated as defect density measured defined as total number of defects present in the software divided by the size of the software. The paper proposes a fuzzy logic based model to predict per phase software defect density. The model uses 3 relevant software metrics per SDLC phase. Defect density prediction is a useful measure, which indicates the critical modules of the project and helps software teams to plan their resources in an efficient manner. The proposed model results are better in comparison with existing literature in the same domain when compared using MRE performance measure on 20 project dataset.