PM2.5 Concentration Prediction Model Based on Chaotic Genetic Neural Network

Abstract

In order to more accurately predict atmospheric PM2.5 concentration and improve prediction accuracy, this paper proposes a method combining phase space reconstruction and chaotic genetic neural network to predict PM2.5 concentration. The mutual information method and GP algorithm are selected to establish the delay time t and the embedding dimension d. The chaotic characteristics of the PM2.5 concentration time series are determined by phase space reconstruction. In order to avoid the problem that BP neural network is easy to fall into the local solution, the genetic algorithm is used to optimize the parameters of chaotic neural network, and the chaotic genetic neural network prediction model is constructed. The model was applied to the prediction of PM2.5 concentration in Beijing, and the PM2.5 concentration in different periods of 100, 500 hours was predicted. The prediction results show that the method can reflect the trend of PM2.5 concentration change in different terms, and can provide a scientific reference for air pollution control.