Forecast of air pollution index based on BP neural network

According to meteorology, the affecting factors which influence the air pollution index are chosen by analyzing several kinds of contaminations and climatic factors. The optimal network structure is determined by making an integrated survey of the approximation capability and the generalization of the network. The fitting and forecasting results indicate that compared with the normal step wise regress method, the forecast precision with the BP method is improved. Moreover, we can gain a highly precise forecast result in the trend of increasing dramatically and declining sharply. The experiment results show that this method can be put into practical use.