The Prediction of PM2.5 Value Based on ARMA and Improved BP Neural Network Model

Abstract

According to the value of PM2.5 in Langfang city in 2015, we propose a new prediction technique based on ARMA and improved BP neural network to forecast the PM2.5 concentrations. In order to prove the accuracy of prediction of PM2.5 concentration of the combined model, Two models considered in the study:ARMA+BP neural network combined model and ARMA+improved BP neural network combined model. The study showed that compared with the ARMA+BP neural network combined model, ARMA+improved BP neural network combined model can better predict the value of PM2.5 and the prediction error is relatively small.