Research On Quantitative Portfolio Strategy

The Copula function can decompose the joint distribution into unrelated edge distributions and a Copula function that can reflect the correlation of edge distributions.Because of this feature,Copula functions are widely used in risk management.In this paper,we selects data from Shenwan Level 1 Industry Index of China's stock market and we uses ARMA(1,1)-gjrGARCH(1,1)model to fit the time series to eliminate sequence autocorrelation and heteroskedasticity,and obtain standardized residual sequences.Then we use the Extreme-Value-Eheory(EVT)to set a threshold for the residual,and use the Generalized Pareto Distribution(GPD)to fit the tail distribution to obtain the extreme distribution of the residual.Extremum theory can describe time-series thick-tailed features well.On the other hand,t-Copula pays attention to the tail dependence,so we choose this function to model the Extreme theory distribution and estimate the corresponding parameters by the maximum likelihood method.Monte Carlo simulations were used to generate residual sequences with Copula dependencies,and VaR and CVaR were calculated from the simulated data.The mean-VaR and mean-CVaR models were constructed to analyze the effective frontier of the portfolio.Based on this,a quantitative investment strategy was designed,and actual market data was used for backtesting to verify the validity of the model.The Copula function can decompose the joint distribution into unrelated edge distributions and a Copula function that can reflect the correlation of the edge distributions.The structure is clear,so it has a wide range of applications in risk management.This paper selects data from 4 plates of China's stock market and uses ARMA(1,1)-gjrGARCH(1,1)model to fit the time series to eliminate sequence autocorrelation and heteroskedasticity,and to obtain standardized residual sequences.Then use the extreme value theory(EVT)to set a threshold for the residual,and use the generalized Pareto distribution(GPD)to fit the tail distribution to obtain the extreme distribution of the residual.Extremum theory can describe time-series thick-tailed features well.t-Copula pays attention to the tail dependence,so we choose this function to model the mechanism distribution and estimate the corresponding parameters by the maximum likelihood method.Monte Carlo simulations were used to generate residual sequences with Copula dependencies,and VaR and CVaR were calculated from the simulated data.The mean-VaR and mean-CVaR models were constructed to analyze the effective frontier of the portfolio.Based on this,a quantitative investment strategy was designed,and actual market data was used for backtesting to verify the validity of the model.