Linguistic fuzzy modeling approach for daily peak load forecasting

Electric load forecasting is absolutely necessary for effective power system planning and operation. Among existing methods for load forecasting, artificial neural network (ANN) and support vector regression (SVR) have shown good forecasting performance. However, ANN and SVR have two drawbacks: 1) black box problem that we don´t know how the prediction models work, 2) high model´s complexity by using many inputs such as type of day indicators (calendar information).