| In today’s world,economic globalization is undergoing rapid development,providing a strong impetus for the growth of the world economy and the general trend of economic and financialization.The stock market plays a leading and pivotal role in the entire modern market system.However,the stock market is uncertain,the stock price is complex and changeable,and various micro and macro factors will affect investors’ judgment on the stock value,making investors The expected rate of return is not achieved.Therefore,stock market research has been widely concerned.Many scholars and industry elites at home and abroad have proposed many prediction methods to improve the accuracy of stock price prediction.However,the high noise and nonlinear stock market makes their research model based on traditional methods.The prediction effect is stagnant.How to break the bottleneck and establish a model with higher accuracy of stock price prediction has great theoretical and practical value for both academia and industry.Based on this,this paper will propose a new forecasting model framework based on the Standard & Poor’s 500 Index.The US stock market is a developed stock market.The market is relatively mature and stable,and the complexity is relatively low.The S&P 500 is the most influential index in the US stock market and can better represent the overall trend of blue chips in US stocks.At the same time,studying the dynamic forecasting model of the US stock market has certain reference significance for the more complex Chinese stock market framework prediction model.Therefore,this paper believes that the selection of the S&P 500 as a forecasting target is representative.Oil has become one of the most important strategic materials in the world,and oil prices have a major impact on the economies of all countries.According to the theory of effective market,behavioral finance theory and actual data,crude oil price and stock price are closely related.Based on this,this paper proposes a hybrid model of stock price forecasting assisted by oil price forecasting.The empirical analysis is divided into two parts.The first part is based on WTI crude oil.WTI(West Texas Intermediate)crude oil is the light crude oil of West Texas,and is known as the international crude oil market with British Brent crude oil and Dubai crude oil in the Middle East.The three major market pricing standards.This paper selects the WTI crude oil commodity day data from January 2,1990 to June 22,2018,including the five indicators of opening price,highest price,lowest price,closing price and trading volume.Innovatively proposed independent component analysis-deep learning prediction model,and found that this model can greatly improve the prediction effect.The second part combines the highest and lowest prices of crude oil forecast with the data of the Standard & Poor’s 500 Index according to the proposed oil price forecasting auxiliary stock price forecasting model,and predicts the daily high and low price of the stock price,and the relevant evaluation indicators.As a basis for judgment,it is found through vertical and horizontal comparison that this model has greatly improved the common non-hybrid prediction model,indicating the necessity of predicting crude oil price and using it for stock price prediction.In general,the forecasting effect for the highest price is better than the forecasting effect for the lowest price.The time period in which the forecasting effect is optimal is before the subprime mortgage crisis,followed by the subprime mortgage crisis,and the subprime mortgage crisis is the worst. |
