Logo

statsmodels.sandbox.regression.gmm.DistQuantilesGMM

class statsmodels.sandbox.regression.gmm.DistQuantilesGMM(endog, exog, instrument, **kwds)[source]

Estimate distribution parameters by GMM based on matching quantiles

Currently mainly to try out different requirements for GMM when we cannot calculate the optimal weighting matrix.

Methods

calc_cov_params(moms, gradmoms[, weights, ...]) calculate covariance of parameter estimates
calc_weightmatrix(moms[, method, wargs]) calculate omega or the weighting matrix
cov_params(**kwds)
fit([start]) Estimate the parameters using default settings.
fitgmm(start[, weights]) estimate parameters using GMM
fititer(start[, maxiter, start_weights, ...]) iterative estimation with updating of optimal weighting matrix
fitonce([start, weights, has_optimal_weights]) fit without estimating an optimal weighting matrix and return results
fitstart()
get_bse([method]) method option not defined yet
gmmobjective(params, weights) objective function for GMM minimization
gradient_momcond(params[, epsilon, method])
jtest() overidentification test
momcond(params) moment conditions for estimating distribution parameters by matching
momcond_mean(params) mean of moment conditions,

Attributes

bse standard error of the parameter estimates

Previous topic

statsmodels.sandbox.regression.gmm.NonlinearIVGMM.momcond_mean

Next topic

statsmodels.sandbox.regression.gmm.DistQuantilesGMM.calc_cov_params

This Page