Software Defect Estimation Using Machine Learning Algorithms
Burcu Yalçıner, Merve Özdeş · 2019 4th International Conference on Computer Science and Engineering (UBMK) · 2019
Software Engineering is a comprehensive domain since it requires a tight communication between system stakeholders and delivering the system to be developed within a determinate time and a limited budget. Delivering the customer requirements include procuring high performance by minimizing the system. Thanks to effective prediction of system defects on the front line of the project life cycle, the project's resources and the effort or the software developers can be allocated more efficiently for system development and quality assurance activities. The main aim of this paper is to evaluate the capability of machine learning algorithms in software defect prediction and find the best category while comparing seven machine learning algorithms within the context of four NASA datasets obtained from public PROMISE repository [12]. All in all, the results of ensemble learners category consisting of Random Forests (RF) and Bagging in defect prediction is pretty much its counterparts.