Intellectualization projection pursuit regression model used in the water demand forecasting

Abstract

In order to solve some actual problems in water demand forecasting such as nonlinearity, high-dimension and so on, though analyzing traditional genetic algorithm (GA) and particle swarm optimization algorithm (PSO)'s advantages and disadvantages in the optimization process, the PSO-GA hybrid algorithm this is researched and carried on the coupling with the projection pursuit regression model based on Hermit multinomial to optimize the best projection direction. Thus an intellectualization projection pursuit regression model is established and used in the water demand forecasting of Tangshan City. It is showed the model is reliable, the forecasting precision is quite high, and it is applicable in the water demand forecasting. © 2012 Springer-Verlag.