International Oil Price Volatility Analysis And Prediction Method Research

As one of the most important strategic resources in the world today,oil plays a pivotal role in the development of modern economy and modern industry.In modern society,from production to life,almost all activities are closely related to oil.This means that every sharp fluctuation in oil prices will have a profound impact on the world economy and even life.The oil crisis that has occurred has greatly confirmed this theory.With the rapid development of China's economy,China's consumption of oil has grown rapidly,and oil production is far from meeting demand.The contradiction between supply and demand has become increasingly prominent.Therefore,oil imports can only be used to alleviate this contradiction.Since China changed from a net oil exporting country to a net oil importing country in 1993,the dependence on foreign oil has increased year by year,reaching a height of 67.4%by 2017.BP World Energy Outlook 2018 shows that in the next few decades,oil will remain an important part of the global energy structure,and China's demand for oil will rise further,meaning that oil dependence may rise.Excessive dependence on foreign oil makes the security of China's energy supply have great hidden dangers.The changes in international oil prices will have a great impact on China's economy and people's production and life.Therefore,analyzing the fluctuation law of international oil prices and predicting the future trend of international oil prices has certain theoretical value and practical guiding significance for China's economic development.This paper takes the spot price of Brent crude oil as the research object,analyzes the fluctuation characteristics and predicts the trend of oil price.Firstly,the fluctuation characteristics of Brent crude oil spot price are analyzed on the whole sample interval and subsample interval respectively to understand whether the fluctuation characteristics of oil price are different in different sample intervals.On the whole interval,the linear and nonlinear GARCH family model are used to characterize the fluctuation features of Brent crude oil spot price.It is concluded from the overall interval that the EGARCH model has the best fitting effect on the volatility features.Therefore,the EGARCH model is used to describe the fluctuation characteristics of the spot price of Brent crude oil in the subsample interval.The empirical results show that the Brent crude oil spot price log rate series has volatility agglomeration,"spike thick tail" characteristics,long memory and asymmetry,both in the overall interval and subsample interval.Although the fluctuations in the spot price of Brent crude oil are asymmetric in different sample intervals,the impact of positive impact on oil price is large,or the impact of negative impact on oil price is large,it needs to be depending on the specific situation.In the interval where the fluctuation is relatively stable,the impact of the positive impact on the current volatility is greater than the impact of the negative impact of the same size on the current volatility.In the interval where the fluctuation is more severe,the impact of the negative impact on the current volatility is greater than the impact of the positive impact of the same size on the current volatility.After studying the fluctuation characteristics of the spot price of Brent crude oil,then predicted its trend.In this paper,the spot price of Brent crude oil is predicted by using the single model ARIMA model,the combined model ARIMA-SVM model,the tune.svm function and the genetic algorithm to optimize SVM model parameters of the ARIMA-SVM model.According to the empirical study of the volatility characteristics,the Brent crude oil spot price series has both linear fluctuation characteristics and nonlinear fluctuation characteristics.The use of the ARIMA model to predict it has certain limitations and does not adequately characterize all the information on the Brent crude oil spot price series.The combined model uses the ARIMA model to extract the linear part information of the Brent crude oil spot price series,and uses the SVM model to characterize the residual part,and at the same time characterizes the linear and nonlinear features,so it has higher prediction accuracy.Under different parameters,the prediction accuracy of the SVM model will be very different.After optimizing the parameters of the SVM model,the prediction accuracy is obviously improved.So the final conclusion isthat the combined prediction model ARIMA-SVM model has higher prediction accuracy than the single prediction model ARIMA model.The prediction accuracy of the ARIMA-SVM model is improved by optimizing the SVM model parameters,and the ARIMA-SVM model based on the genetic algorithm to optimize the SVM model parameters has the best prediction performance.Finally,on the basis of summarizing the full text,the shortcomings of this paper are discussed and the future research is prospected.Based on the research of international oil prices by domestic scholars and foreign scholars,this paper attempts to make two innovations:(1)Research perspective.The fluctuation characteristics of international oil prices are analyzed from the perspective of the overall sample interval and the subsample interval.The subsample interval is divided into a section where the price fluctuation is relatively stable and a section where the price fluctuation is relatively severe,the main research is on the similarities and differences of the fluctuation characteristics of international oil prices in different fluctuation ranges.(2)Prediction model.The genetic algorithm was used to optimize the ARIMA-SVM model and applied to the prediction of international oil prices,which achieved good prediction results.