statsmodels.sandbox.regression.gmm.GMMResults.calc_cov_params¶

GMMResults.calc_cov_params(moms, gradmoms, weights=None, use_weights=False, has_optimal_weights=True, weights_method='cov', wargs=())[source]¶

calculate covariance of parameter estimates

not all options tried out yet

If weights matrix is given, then the formula use to calculate cov_params depends on whether has_optimal_weights is true. If no weights are given, then the weight matrix is calculated with the given method, and has_optimal_weights is assumed to be true.

(API Note: The latter assumption could be changed if we allow for has_optimal_weights=None.)