Application of Coupling Model with Neural Network and Projection Pursuit Based on Partial Least-Squares Regression to Water Resources Carrying Capacity Forecasting

Abstract

The method of partial least-squares regression can effectively deal with the problems of multicollinearity among independent variables", "but can not ideally solve the complicated problems of nonlinearity between dependent variables and independent variables. The method of coupling model with neural network and projection pursuit is an ideal tool to deal with the problem of nonlinearity, and it is very steady, but can not ideally solve the problems of multicollinearity among independent variables. The paper combines the two methods to establish the method of coupling model with neural network and projection pursuit based on partial least-squares regression for forecast water resources carrying capacity. the results of forecasting indicate that the combination is superior to either of them, the model was found to be able to give satisfactory effect.