Latent Volatility Modeling and Bayesian Analysis of stochastic Volatility of Intraday Data of Tehran Stock Exchange Index Based on Markov Monte Carlo Chain

In this study, latent volatility modeling and Bayesian analysis of stochastic Volatility of intraday data of Tehran Stock Exchange index based on Markov Monte Carlo chain in uncertainty conditions (downward crisis of stock market index) have been developed. The method of the current research is a correlational description. For this purpose, at first, the distribution of the logarithm of the squared return as a measure of the realized volatilities was simulated using the stochastic Volatility model to obtain the latent volatilities, and then by using the hybrid MCMC-Copula model, the parameters affecting the stochastic Volatilities were identified and estimated in the training phase. Finally, using the results obtained from the training phase, in the test phase, the comparison of Copula and GARCH models was done. The results showed that the Copula Gumble, Galambos, Joe, Clayton and Frank provide similar and lower MSE and RMSE indices than the GARCH base model, and therefore the model based on copula provides the possibility of serial dependence in the latent volatility process. The findings of the current research can be useful for financial and investment companies for portfolio management and portfolio management in different conditions of market volatilities in order to achieve the investor's goals and increase the value of the portfolio.