The Study and Application of Two Combined Forecast Models Used in the Three Gorges Reservoir Long-Term Runoff Forecast

Abstract

Long-term hydrological forecast is an integral part of hydrological forecast. Its level of accuracy plays an important role in the optimal allocation of water resources. In this paper, we bring out a combined forecast model based on Support Vector Machines (SVM) and entropy weight and apply it to the Three Gorges Reservoir runoff prediction. Compared to the Nearest Neighbor Bootstrapping Regressive model (NNBR), Mean Generating Function (MGF) and Automatic Regressive model (AR), the combined model is better than the individual model. It provides a reliable prediction method for long-term runoff forecast.