The Application Of Time Series Models In Stock Market

The stock market is unpredictable,and while it brings high returns for investors,it is also often accompanied with high risk.In order to control the investment risk and master the stock market effectively,it is of great significance to make research on the stock market and risk measure of stock rate of return.As one of the main methods in the field of short-term prediction,the theory and method of time series analysis have been widely applied to every aspect of social life.As typical financial time series data,stock prices or stock returns often present characteristics of conditional heteroscedasticity,peak and thick tail,asymmetry and others.The family of GARCH models can deal with characteristics of conditional heteroscedasticity and asymmetry effectively,but it is not effective when dealing with features of peak and fat-tailed.Quantile regression is introduced by Roger Koenker in 1978,quantile regression itself has excellent properties.First of all,it doesn't have any requirements on the distribution of random error,and it is robust on outliers.Second,it can measure the influence of dependent variables on the tail of distribution and get more comprehensive information.Therefore,it is of great significance and application values to make research on stock price or risk measure of stock rate of return by combining quantile regression and time series models.This paper mainly studies the application of time series models on the stock market.The main contents are as follows:first,The relevant definition of quantile regression method and its properties is discussed.It shows that quantile regression has more advantages than least squares regression at some aspect.Second,the models of time series are discussed.It includes models of stationary time series and models of conditional heteroscedastic.Third,Risk measures of Shanghai index rate of return based on the model QR-t-GARCH(1,1)is studied.The definition of value at risk(VaR)and the its methods of calculation is included.It is discovered that it is better to adopt model of quantile regression to study the effect of risk measure on Shanghai index rate of return.This method can be applied to other financial markets.Fourth,Short-term prediction analysis of Shanghai composite index based on models of time series is studied.It is discoverd that it is better to adopt model of ARMA(3,3)-GARCH(1,1)which residuals obey t distribution to study short-term prediction on Shanghai composite index.This model can furtherly be used on short-term prediction of Shanghai composite index,and the results of prediction can provide reference suggestions for relevant stock investors.