Research On The Measurement Of Time Series Risk Of Securities Based On Series Quantile Regression

Volatility is a key variable for the financial market to measure risk,and its accurate measurement is of great significance for preventing financial risks and promoting the sound operation of the economy.In recent years,research on the construction of applicable topics such as predictive models and reasonable measurement of financial risks has never stopped,but due to the characteristics of financial data such as spikes and thick tails,obedience to skewed distribution,and often accompanied by high volatility,high risk and other fluctuations This kind of phenomenon brings great difficulty to model construction and other issues.It is difficult to give an accurate and objective description of volatility using traditional time series models and parameter estimation methods.In addition,with the rapid development of computer technology,the acquisition of high-frequency financial data has become easier and easier,and it often contains more useful financial information than low-frequency data.Based on this,how to realize the prediction of high-risk and high-frequency fluctuations in the financial market and give a reasonable measure of financial risks is an issue that requires our in-depth discussion.It has important theoretical and application values.China's GEM market has the structural advantages of special industries,but also has market characteristics such as high volatility and high risk,which is the focus of current academic research.In view of the fact that the EGARCH model can well characterize the asymmetry of financial risks and other advantages,while considering that quantile regression estimation has many advantages such as unlimited distribution and strong applicability,this article attempts to construct The EGARCH model of a series of quantile regression estimation methods,combined with the realized volatility,carries out the realized volatility analysis and risk measurement estimation of high-frequency data in China's GEM market.The main research contents include:1.The model analysis of the combination of the series quantile(quantile(QR),composite quantile(CQR)and weighted composite quantile(WCQR))regression estimation method and the EGARCH model is provided,including: Parameter estimation derivation and estimation formulas to explore the consistency and robustness of the estimation methods;give the selection of the optimal weights of the WCQR estimation method;introduce the identification of jump and continuous fluctuations of realized volatility under high-frequency data Method,etc.2.For the closing data of China's Chi Next Composite Index,verify and analyze the asymmetry of the Chi Next market and confirm the applicability of the EGARCH model;and use realized volatility to construct a series of quantile regression estimation methods.The EGARCH model is realized.By calculating 6 commonly used loss functions,the advantages and disadvantages of the EGARCH model under different estimation methods are given.3.Mainly aiming at the situation where the fitting effect of the composite quantile estimation of the EGARCH(1,1)model is not ideal,the model is further modified to consider the impact of the jump phenomenon on the return to different degrees,and the jump volatility time series Jt The constructed EGARCH-J model with jump component was constructed,and the parameter estimates of the realized EGARCH-J model under the series quantile regression estimation method were obtained;and through comparison of the same loss function,it was judged that it is more suitable for China's gem market Volatility fit model.The research results show that the EGARCH model based on the series quantile regression estimation method constructed in this paper is reasonable and effective.Jumping volatility as an explanatory variable helps improve the fit of volatility models.Considering the impact of jumping volatility on yield,weighted compound quantile regression estimation method has the best fitting effect on volatility,which is better than compound The quantile regression estimation method is better than the quantile regression estimation method,and it makes up for the shortcomings of the traditional methods such as the least squares method.It further proves that the research method in this paper is suitable for high-risk financial data and high-frequency fluctuation prediction.Superiority.The research methods and related models in this paper have achieved effective estimation of China's GEM market risk measurement and volatility,especially for many large jumps in financial risk.The model established in this paper gives a detailed analysis of market fluctuations under different degrees of shock And accurate simulation,thus revealing the changing pattern of high-risk fluctuations in China's GEM market,and providing theoretical support for reasonable risk aversion.