ENH/BUG: allow fit keyword ridge_factor in newton

[josef-pkt]

model.fit(method="newton", ridge_factor= 1e-5)

issues a warning about unused kwarg, and ridge_factor is not transmitted to the newton optimizer.
I wanted to try out fit with a large ridge factor to force convergence (to a possibly regularized solution)

I'm not sure whether this ever worked or was a refactoring victim.

AFAICS, it's because ridge_factor is explicit keyword and not in kwargs in signature, base.optimizer

        def _fit_newton(f, score, start_params, fargs, kwargs, disp=True,
                maxiter=100, callback=None, retall=False,
                full_output=True, hess=None, ridge_factor=1e-10)