Influence of Noise to Chaotic Time Series Prediction in Environment

When constructing chaotic time series prediction model involving noise, noise would influence the chaotic characteristics of the time series and lower the precision and generalization ability of the model. This paper applies wavelet transform to make de-noise on gas emission time series. It uses state space reconstruction theory to determine the embedding dimension and delay time. In the constructed phase space, build BP artificial neural network model to make prediction. Prediction model constructed after wavelet de-noises achieves good prediction performance.