Research on Crude Oil Price Forecasting Based on MLR-SVR Decomposition Integrated Model

Crude oil prices often show the characteristics of random changes. Mastering the changing trend of international crude oil prices can not only affect China's political stability and economic development but also help investors control the risks of crude oil. In this thesis, a mixed crude oil price prediction model with multiple linear regression (MLR) and support vector regression (SVR) was established by us. The model was applied to predict the international crude oil futures price WTI on the New York Mercantile Exchange and compare the forecast results with the predictions of multiple linear regression and support vector regression. The results show that the proposed hybrid model has high fitting and prediction accuracy, among which, the MAE is 1.12, which is 2.7% and 61.2% higher than that of the single model, respectively.