A Novel BP Neural Network with Wavelet Transform Inputs for Air Quality Index Prediction

Abstract Air quality index prediction research is of great significance to people’s life. Various prediction methods are used to pursue its accurate prediction. A novel BP neural network with wavelet transform inputs for air quality index prediction is proposed in this paper. The proposed method introduces wavelet transform in the front of the neural network. It decomposes the original time series signal into a linear superposition of the main signal and some noise signals. After establishing the predictive BP neural network for each decomposed signal, the air quality index is restored based on the linear additivity of wavelet decomposition. The simulation experiments show that the prediction accuracy of proposed method is high. The relative error of long-term prediction for 50 consecutive days is only 4.3994% for Xuzhou City’s air quality index. The research work provides a useful reference for air quality index prediction.