I'm trying to estimate some parameters using the GMM approach (Generalized Method of Moments, not Gaussian Mixture Model). I was hoping to use the package statsmodels if possible. The official description is listed here. However I got confused about where to specify the moment conditions. My model is not linear and does not use IV, so I cannot use NonlinearIVGMM and such. The "tutorials"/examples I found are only for models with IVs. I wonder if there are any examples to showcase how to use the more general GMM. Thanks.
1 Answer
Yes! You could see
Issue with using statsmodels.sandbox.regression.gmm.GMM
and
to have a good picture of the function. Basically you have to declare a class using GMM function from statsmodels and then a subclass defining your moment conditions and then, run.
