Prediction of Air Quality Index Based on Wavelet Transform Combination Model

In the prediction of air quality index, atmospheric space is a large and complex system with many interfering factors, which leads to the nonlinearity, time-variability and uncertainty of AQI time series. This paper takes the data provided by Nanjing environmental protection testing center as the research object.Firstly, wavelet analysis was used to preprocess the AQI time series.Then ARIMA model and LS-SVM model are used to predict the linear autocorrelation part and the nonlinear part. PSO algorithm is used to optimize the parameters of LS-SVM model. Simulation experiments show that, after selecting the appropriate wavelet decomposition function and parameters of the combined model, the combined model does reduce errors compared with the traditional ARIMA model and LS-SVM model, and has higher accuracy. It provides a new way to forecast the actual air quality index.