A phase-wise fault prediction using soft computing

Abstract

Software testing is performed to increase quality of the software and also increase the customer satisfaction. In today’s scenario, most of the work is dependent upon software and need of quality software in minimum time. If the fault is not found in the early development phase is having more chance of failure rate in a later phase. Since the failure data or test data are not available in requirement, design, coding and testing phase. For this goal, the phase-wise approach is proposed using soft computing to find the fault in software development life cycle (SDLC) phase. In proposed approach, top reliability-based software metrics/quality attributes for fault prediction is selected from the phases of SDLC. In the proposed approach, four fuzzy inference systems are designed for SDLC phases. The given result from the proposed approach is validated with ten real software datasets from promise datasets. The predicted fault results are near to actual dataset fault. The result from proposed approach is useful for a software developer team, design team, and testing team to plan and take decision for their respective phases of development.