KOMPARASI ALGORITMA DENGAN PENDEKATAN RANDOM UNDERSAMPLING UNTUK MENANGANI KETIDAKSEIMBANGAN KELAS PADA PREDIKSI CACAT SOFTWARE

Testing is a process that becomes a standard in producing quality software. In predictions of software defects, prediction errors are very bad. Incorrect and inappropriate data sets result in inaccurate prediction results will be affect the software itself. This study aims to overcome the problem of class imbalance with the software defect prediction data set, through the Random Undersampling (RUS) data level approach by taking several algorithms namely Naive Bayes (NB), J48 and Random Forest (RF) which aims to compare the accuracy level highest so that maximum results are obtained in the process of predicting software defects. From the results of this study it can be found that to overcome class imbalances using the Random Undersampling level data approach to predict software defects, the highest level of accuracy is obtained by the Random Forest algorithm with an accuracy rate of 71.932%