rethink base Estimator

[kjhall01]

• some methods .fit on continous observations data, and are capable of producing probabilistic forecasts.
• predict_proba takes the number of outputs from the number of outputs on the training data
• methods trained on continuous data expect only a single output for predict_proba - but that is an issue for tercile forecasts

[kjhall01]

we can also add broadcasting along higher dimensions , if implementing by stacking x/y/...higher dimensions along a 'point' axis.

[kjhall01]

also: remove extraneous transposing / rechunking / appending

[kjhall01]

you can remove Nan calculations by stacking and dropping missing points externally. multioutput is implemented for predict and predict proba. transposing/rechunking has been addressed to an extent.