Full sklearn compliance

[falkben]

So that it passes the estimator test.

First thing to do is to make the model pickleable.

[adam2392]

Hi @falkben was there ever any update on pickling the forest classifiers to be ran through something like joblib?

[falkben]

No, I'm sorry, I don't have an update on that. My guess is that it won't work with joblib's dump method. Have you tried it?

[pepetikesavasiddhardha]

@falkben I tried to pickle it with joblib and pickle libraries but it was not working, its showing an error like this
TypeError: cannot pickle 'pyfp.fpForest' object
So can u suggest me how can i save this method? Is there any way to do that?

[adam2392]

@pepetikesavasiddhardha currently the tree models are implemented inline with sklearn in https://github.com/neurodata/scikit-tree