DCCGARCH  过时了   线性相关系数 有太大局限性    用动态copula模型

[parameters, loglikelihood, Ht,...]=dcc_mvgarch(r,1,1,1,1);
var1 = squeeze(Ht(1,1,:));  %  column vector of variances
var2 = squeeze(Ht(2,2,:));  %  column vector of variances
cov12 = squeeze(Ht(1,2,:)); %  column vector of covariances
corr12 = cov12./sqrt(var1.*var2);

[DCC_parameters,DCC_LL,DCC_Ht]=dcc_mvgarch(data,1,1,1,1);
Estimating GARCH model for Series 1
Warning: Options LargeScale = 'off' and Algorithm = 'trust-region-reflective' conflict.
Ignoring Algorithm and running active-set method. To run trust-region-reflective, set
LargeScale = 'on'. To run active-set without this warning, use Algorithm = 'active-set'.
> In fmincon at 412
  In fattailed_garch at 198
  In dcc_mvgarch at 82
Estimating GARCH model for Series 2
Warning: Options LargeScale = 'off' and Algorithm = 'trust-region-reflective' conflict.
Ignoring Algorithm and running active-set method. To run trust-region-reflective, set
LargeScale = 'on'. To run active-set without this warning, use Algorithm = 'active-set'.
> In fmincon at 412
  In fattailed_garch at 198
  In dcc_mvgarch at 82


Estimating the DCC model
Warning: Options LargeScale = 'off' and Algorithm = 'trust-region-reflective' conflict.
Ignoring Algorithm and running active-set method. To run trust-region-reflective, set
LargeScale = 'on'. To run active-set without this warning, use Algorithm = 'active-set'.
> In fmincon at 412
  In dcc_mvgarch at 98

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
   Diagnostic Information

Number of variables: 2

Functions
Objective:                            dcc_mvgarch_likelihood
Gradient:                             finite-differencing
Hessian:                              finite-differencing (or Quasi-Newton)

Constraints
Nonlinear constraints:             do not exist

Number of linear inequality constraints:    1
Number of linear equality constraints:      0
Number of lower bound constraints:          2
Number of upper bound constraints:          0

Algorithm selected
   medium-scale: SQP, Quasi-Newton, line-search


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
End diagnostic information


                                Max     Line search  Directional  First-order
Iter F-count        f(x)   constraint   steplength   derivative   optimality Procedure
    0      3       2092.4    -0.009998                                         
    1     11      2088.75     -0.01937       0.0313         -702    1.76e+003   
    2     14       2088.6            0            1          -75    1.67e+003   
    3     18      2081.78     -0.00467          0.5         -771          599   
    4     22      2081.24    -0.002335          0.5         -123    1.37e+003   
    5     25      2080.53   -1.11e-016            1   -1.89e+003     1.6e+003   
    6     36      2080.53   -0.0006755      0.00391         -107          686   
    7     39      2080.11   -0.0007109            1         -202           99   
    8     42      2080.11   -0.0007696            1        -55.3         8.15   
    9     45      2080.11   -0.0007977            1        -7.97         1.19   
   10     48      2080.11   -0.0007944            1        -1.02       0.0392   

Local minimum possible. Constraints satisfied.

fmincon stopped because the size of the current search direction is less than
twice the default value of the step size tolerance and constraints were
satisfied to within the selected value of the constraint tolerance.

<stopping criteria details>

No active inequalities.




从估计的过程来看，dcc——mvgarch命令是不需要做单变量garch估计的吧，因为从上面的结果来看该命令首先做了一个garch模型，然后再做dcc估计的，不知道是不是这样的？做过的来讨论一下！谢谢！