Natural gas prediction model based on wavelet transform and BP neural network

In order to improve the prediction accuracy of natural gas load, a decomposition-prediction-reconfiguration natural gas forecasting model based on wavelet transform and BP neural network is proposed in this paper. Firstly, we used the Mallat fast algorithm to decompose the sample sequence of natural gas load to eliminate the influence of noise factors on prediction; Secondly, for the decomposed low-frequency signal, the BP neural network is used to fit the prediction; finally, we superimposed the noise signal on the wavelet reconstruction to obtain the final prediction result. Using the gas load data of Wuhan in 2014 and 2015 to test the model, it can be concluded that the predicted relative error percentage of the natural gas forecasting model based on wavelet transform and BP neural network is 2.79%, which is better than the traditional forecasting model and have stronger robustness.