- class statsmodels.sandbox.regression.gmm.GMM(endog, exog, instrument, k_moms=None, k_params=None, missing='none', **kwds)¶
Class for estimation by Generalized Method of Moments
needs to be subclassed, where the subclass defined the moment conditions momcond
endogenous variable, see notes
array of exogenous variables, see notes
array of instruments, see notes
number of moment conditions, if None then it is set equal to the number of columns of instruments. Mainly needed to determine the shape or size of start parameters and starting weighting matrix.
this is mainly if additional variables need to be stored for the calculations of the moment conditions
The GMM class only uses the moment conditions and does not use any data directly. endog, exog, instrument and kwds in the creation of the class instance are only used to store them for access in the moment conditions. Which of this are required and how they are used depends on the moment conditions of the subclass.
Options for various methods have not been fully implemented and are still missing in several methods.
TODO: currently onestep (maxiter=0) still produces an updated estimate of bse and cov_params.
calc_weightmatrix(moms[, weights_method, ...])
calculate omega or the weighting matrix
fit([start_params, maxiter, inv_weights, ...])
Estimate parameters using GMM and return GMMResults
fitgmm(start[, weights, optim_method, ...])
estimate parameters using GMM
fitgmm_cu(start[, optim_method, optim_args])
estimate parameters using continuously updating GMM
fititer(start[, maxiter, start_invweights, ...])
iterative estimation with updating of optimal weighting matrix
from_formula(formula, data[, subset, drop_cols])
Create a Model from a formula and dataframe.
objective function for GMM minimization
gmmobjective_cu(params[, weights_method, wargs])
objective function for continuously updating GMM minimization
gradient_momcond(params[, epsilon, centered])
gradient of moment conditions
mean of moment conditions,
After a model has been fit predict returns the fitted values.
score(params, weights[, epsilon, centered])
score_cu(params[, epsilon, centered])
set the parameter names in the model
Create identity matrix for starting weights
Names of endogenous variables.
Names of exogenous variables.