SOFTWARE DEFECT PREDICTION USING ENSEMBLE OF MACHINE LEARNING TECHNIQUES

Abstract

Now a days research on software defect prediction has attracted many researchers because it helps in creation of successful software. Additional advantage is that it helps in reduction of the software development cost and facilitates procedures to identify the reasons for determining the percentage of defect-prone software in future. For specific types of machine learning, there is no conclusive evidence that will be more efficient and accurate in predicting software defects. Some of the previous related work, however, proposes the learning techniques of the ensemble as a more precise alternative. This work introduces the resample technique with three types of ensemble learners; boosting, bagging, stacking and voting using four base learners on different versions of same dataset repository provided in the PROMISE repository. Results indicate that accuracy has been improved using ensemble techniques more than single leaners.