getvarmatrix_ms_ar.m

matlab自回归马尔可夫转换模型仿真估计与预测

% Function for calculating standard errors of MS_Regress_Fit
%
% The calculation (aproximation) of the first and second derivative of the likelihood
% function is done by one side finite differences method
%
% Four methods for the calculation of the covariance matrix were implemented here:
% (1) Using second partial derivatives
% (2) Using first partial derivatives (outer product Matrix)
% (3) Using white covariance matrix 
% (4) Using Newey and West Covariance matrix
% 
% The standard erros from the last two methods are also called robust
% standard errors.
%
% References:
%
% HAMILTON, J., D. (1994). Time Series Analysis. Princeton University Press.
%
% NEWEY, B., WEST, D (1987) 慉 Simple, Positive semidefinite, Heteroskedasticity
% and Autocorrelation Consistent Covariance Matrix