Modeling Of Risk Measurement For Chinese Stock Market Based On Hill Change Point Threshold Selection

The risk measurement of financial market is a very frontier and hot research topic.Many researchers have studied the risk measurement of Chinese stock market.As well know,the stock market is vulnerable to many uncertain factors which cause the fluctuations for the stock price,resulting in different degree of stock market risk.Therefore,this paper try to provide some valuable suggestions for investors to avoid risks by studying the risk measurement of China's stock market.This article arranges as following.Firstly,the finance and risk measurement tools,stochastic volatility models,and two-dimensional Poisson process models are introduced in Chapter 1.The definitions of financial risk measurement tools for VaR and ES are presented in Chapter 2.In Chapter 3,we introduce the idea of risk measurement and threshold selection for the SV-T-POT mode.In the empirical analysis,we select the closing price data of the Shanghai Stock Exchange Index and the Shenzhen Stock Exchange Index from 1996 to 2018 for empirical analysis,model the data based on the SV-T-POT model,and use the Bayesian MCMC method to analyze the parameters of the model.It is estimated that the Hill estimation method of variable point analysis is introduced to quantify the threshold for the standardized residual sequence.The empirical results show that the SV-T-POT model selected by this threshold method can effectively describe the risk measure of China's Shanghai Stock Index and Shenzhen Stock Index.In Chapter 4,we introduce the risk measure of the two-dimensional non-homogeneous Poisson process model based on threshold selection,and select the Guizhou Maotai stock for risk measurement research,introduce explanatory variables into the model parameters,and use the Hill estimation method of variable point analysis.The threshold is selected quantitatively,and the data is re-selected and then modeled.The empirical results show that the model can effectively capture extreme information and obtain more accurate risk values.