Research Article

Unified Quantile Regression Deep Neural Network with Time-Cognition for Probabilistic Residential Load Forecasting

Figure 8

Heat map with AQS metrics for evaluation of four models. QGBRT, QCNN, QLSTM, and our proposed model in X-axis correspond to AQS for different models, while I_QGBRT, I_QCNN, and I_QLSTM represent the AQS improvement ratio compared with the proposed model. Y-axis denotes different residences. Results in each residence are described in the same row, where darker green shades imply a better performance.