The Application Of Threshold Auto-regressive (TAR) Moldel In Stock Price Analysis

In this paper,the nonlinear properties of time series are discussed and studied.The maximum likelihood estimation of parameter estimation of static Threshold Regression Model with individual fixed effect is discussed.The threshold effect is tested by likelihood ratio test,and similar results extended to Single-threshold Model and Double-threshold Model,and use the bootstrap method to repeat the sampling to get its empirical distribution.In addition,the Threshold Auto-regressive Model is used to estimate the parameters of the regression model through the model structure transformation,and the modeling steps of the TAR model are proposed.In order to verify the validity of the method,the R software is used to analyze the stock price data of.Yunnan Baiyao powder(000538).The model is obtained by Hansen test.The feasibility of the model is tested by using the error test model and the model is optimized to establish a threshold Auto-Regressive Model describing the nonlinear nature of the stock price.The results show that the effect of using SETAR model to fit the stock price fluctuation is obvious,which reflects the non-linearity of the sequence.However,because the stock price changes include many factors,the model can only predict the short-term price of the stock,the long-term stock price forecast is not ideal.