Cognitive Inherent SLR Enabled Survey for Software Defect Prediction

Abstract

Any software is created to help automate manual processes most of the time. It is expected from the developed software that it should perform the tasks it is supposed to do. More formally, it should work in a deterministic manner. Further, it should be capable of knowing if any provided input is not in the required format. Correctness of the software is inherent virtue that it should possess. Any remaining bug during the development phase would hamper the application's correctness and impact the software's quality assurance. Software defect prediction is the research area that helps the developer to know bug-prone areas of the developed software. Datasets are used using data mining, machine learning, and deep learning techniques to achieve study. A systematic literature survey is presented for the selected studies of software defect prediction. Using a grading mechanism, we calculated each study's grade based on its compliance with the research validation question. After every level, we have selected 54 studies to include in this study.