Cost-sensitive decision tree ensembles for effective imbalanced classification

Bartosz Krawczyk, Michal Wozniak, Gerald Schaefer

Abstract: Highlights•A novel cost-sensitive ensemble, based on decision trees, for imbalanced classification.•An evolutionary-based simultaneous classifier selection and fusion boost the recognition rate of the minority class.•An analysis of the influence of the cost matrix parameters and data imbalance ratio on the performance of the ensemble.•A ROC-based tuning method of the ensemble parameters.