_fit_minimize requires update of no_hess list: L-BFGS-B and TNC do not use Hessian information (hess)

[pcuestas]

L-BFGS-B and TNC do not use Hessian information (hess). Therefore, these methods should be added to

statsmodels/statsmodels/base/optimizer.py

Line 341 in 23faea3

no_hess = ['Nelder-Mead', 'Powell', 'CG', 'BFGS', 'COBYLA', 'SLSQP']

[josef-pkt]

yes lbfgs are missing but AFAICS we don't have a wrapper for fmin_tnc.

Do you get any problems, exception, ... when running the current code?
I don't understand why this doesn't cause problems when we run the unit tests.

update
no_hess is only used in _fit_minimize and we have few unit tests for it.

[josef-pkt]

we should get at least a smoke test for all methods in _fit_minimize.

[pcuestas]

I do not have any problem when running the current code, only a warning. Sorry, I should have mentioned this. I just wanted to point this out, as I believe it can be confusing for users to get this warning when there shouldn't be one.