Analysis on the Estimators to the OSELM Model

The OSELM model is one of many classifiers for software defect prediction that can detect the specific oversampled input problems (pattern). OSELM was a more suitable model for defect prediction when used with the extremely unbalanced data and noisy like NDP repository datasets. Prior metrics were measured using balanced accuracy and recall. In this study, the estimator metrics for the OSELM model are G-Mean, F1-Score, and Matthews Co-Relation Coefficient