R language garch-dcc model and DCC (MVT) modeling estimation


Original link:http://tecdat.cn/?p=7194

This short presentation illustrates the use ofrThe DCC model of the software package and the use of its method, especially another method for Level 2 DCC estimation in the presence of MVT distribution shape parameters.

The first phase and pass it to dccfit

 cl = makePSOCKcluster(10)

multf = multifit(uspec, Dat, cluster = cl)

Next, the DCC model is estimated.

fit1 = dccfit(spec1, data = Dat, fit.control = list(eval.se = TRUE), fit = multf, cluster = cl)

In order to fit the DCC (MVT) model in practice, either the QML in the first stage is assumed, or the common shape parameters must be estimated together in the stage. In the following example, an alternative method is used to estimate approximate common shape parameters.

The variation of likelihood and shape parameters shows that it can converge to a stable value in only a few iterations.

R language garch-dcc model and DCC (MVT) modeling estimation

The value of the shape parameter indicates that the kurtosis is 1.06. The asymmetric DCC (MVT) model was fitted repeatedly.

xspec = ugarchspec(mean.model = list(armaOrder = c(1,1)), variance.model = list(garchOrder = c(1,1), model = "eGARCH"),  distribution.model = "norm")

The following table shows a summary of the estimation model, with an asterisk next to the coefficient indicating the significance level (* * * * 1%,* 5%, 10%)。

##           DCC-MVN   aDCC-MVN    DCC-MVL   aDCC-MVL      DCC-MVT     aDCC-MVT
## a      0.00784*** 0.00639*** 0.00618***  0.0055***   0.00665***    0.00623***
## b      0.97119*** 0.96956*** 0.97624*** 0.97468***   0.97841***    0.97784***
## g                    0.00439               0.00237                 0.00134
## shape                                                9.63947***    9.72587***
## LogLik      22812      22814      22858      22859        23188         23188

The following chart illustrates some dynamic correlations from different models:

R language garch-dcc model and DCC (MVT) modeling estimation

Terminate cluster object:


R language garch-dcc model and DCC (MVT) modeling estimation


1.Using machine learning to identify changing stock market conditions — Application of hidden Markov model (HMM)

2.R language garch-dcc model and DCC (MVT) modeling estimation

3.R language implementation copula algorithm modeling dependency case analysis report

4.R language copulas and VaR analysis of financial time series data

5.R language multivariate copula GARCH model time series prediction

6.An example of using R language to realize neural network to predict stock

7.Implementation of R language to predict Volatility: arch model and har-rv model

8.How to make Markov Switching Model in R language

9.Matlab uses copula simulation to optimize market risk