Rainfall Forecasting Using Projection Pursuit Regression and Neural Networks

Abstract

Accurate forecasting of rainfall has been one of the most important issues in hydrological research. Due to rainfall forecasting involves a rather complex nonlinear data pattern; there are lots of novel forecasting approaches to improve the forecasting accuracy. This paper proposes a Projection Pursuit Regression and Neural Networks (PPR--NNs) model for forecasting monthly rainfall in summer. First of all, we use the PPR technology to select input feature for NNs. Secondly, the Levenberg--Marquardt algorithm algorithm is used to train the NNs. Subsequently, example of rainfall values in August of Guangxi is used to illustrate the proposed PPR--NNs model. Empirical results indicate that the proposed method is better than the conventional neural network forecasting models which PPR--NNs model provides a promising alternative for forecasting rainfall application.