dask grid search

dask_grid_search.py

from dask_ml.compose import ColumnTransformer
from dask_ml.preprocessing import StandardScaler, DummyEncoder, Categorizer
from dask_ml.model_selection import GridSearchCV

# Dask has slightly different way of one-hot encoding
pipeline = Pipeline(steps=[
    ('categorize', Categorizer(columns=categorical_feat)),
    ('onehot', DummyEncoder(columns=categorical_feat)),
    ('scale', ColumnTransformer(
        transformers=[('num', StandardScaler(), numeric_feat)],
        remainder='passthrough',
    )),
    ('clf', ElasticNet(normalize=False, max_iter=100)),
])

# params same as above
grid_search = GridSearchCV(pipeline, params, cv=3, scoring='neg_mean_squared_error')