The level of defects in a software will always be linear with the quality of the resulting software. In the development process, developers need to predict the level of defects in a software to produce better software. In this study, the Particle Swarm optimization (PSO) method was used to optimize the data at the preprocessing stage, the Random Over Sampling (ROS) method to balance the classes in the dataset and the ensemble technique to maximize the performance of the J48 algorithm. The dataset used in this study uses the PROMISE repository dataset. The results showed that the integration of the PSO+ROS+J48+Bagging algorithm resulted in an average accuracy value of 92.378% and an AUC value of 0.924. This shows that the combination of PSO, ROS and J48 methods with Bagging Technique is feasible to be used as an algorithm to predict the defect level of a software
Copyrights © 2022