1-D vs 2-D y (dependent) array

ContinuumIO / elm

Phase I & part of Phase II of NASA SBIR - Parallel Machine Learning on Satellite Data

http://ensemble-learning-models.readthedocs.io

44 stars 24 forks source link

1-D vs 2-D y (dependent) array #191

Open PeterDSteinberg opened 7 years ago

PeterDSteinberg commented 7 years ago

Address warnings from scikit-learn regarding whether 1-D y array need to be expressed as a 2-D y array with 1 column. Note some regression models predict a y with multiple columns and this will require special treatment, e.g. a mapping estimator names or classes to allowable y shapes

PeterDSteinberg commented 6 years ago

Also see #201 and specify what data structures in that checklist should be supported for y vs X. Currently (post PR #192) y is always a numpy array or y is in a Dataset-like object and extracted by a transformer.

PeterDSteinberg commented 6 years ago

Also consider how sample_weight should be passed (somewhat related to plans for data structures that may be used for y data and how the transformers return a single output or X, y, sample_weight tuple)