RADE: Resource-Efficient Supervised Anomaly Detection Using Decision Tree-Based Ensemble Methods