Research On Crude Oil Price Modeling Based On Wavelet Denoising

Crude oil plays an important role in the world's energy,it not only has the property attributes,but also with a certain political color.In recent decades,with the rapid development of China's economy,the consumption of domestic crude oil resources are also growing.In China's energy resources,crude oil resources are less distributed,far from meeting their needs.So our country can only import crude oil from the international,and imports are also increasing year by year.In this way,the international crude oil price shocks on the healthy development of China's economy can not be ignored,the study of international crude oil price fluctuations is very necessary.Since crude oil price data is collected over time,it can be considered time series data.In view of the fact that the time series model has the advantages of wide application range,low data requirement and high short-term prediction accuracy,it is widely used in model prediction.However,the traditional time series model has the assumption that the data is linear,so it is difficult to capture the hidden nonlinear relationship behind the data for the nonlinear data.The neural network simulates the way in which neurons deal with problems in the human body,and has a good effect on dealing with nonlinear problems.At the same time,it is easier to operate,and can quickly find the optimal solution to the problem.This article selected the January 2,2015 to December 13,2016 Brent oil spot trading price data as the original data sequence.Considering that the high noise in the original data may interfere with the model analysis,the original data is denoised by the wavelet analysis technique before establishing the BP neural network.Then,considering the existence of the convergence speed of BP network itself,it is easy to fall into the local optimal and so on.In this paper,genetic algorithm is introduced to optimize it.The results of empirical analysis show that the denoising of the original data before modeling can improve the prediction accuracy of the model.In addition,based on the establishment of the above model,the genetic algorithm is optimized.To a certain extent,speed up the convergence rate of the model,and further improve the model prediction accuracy.