pretools
Preparation tools for machine learning.
Examples
from pretools.sklearn.estimators import *
from sklearn.datasets import load_boston
from sklearn.pipeline import make_pipeline
X, y = load_boston(return_X_y=True)
model = ModifiedCatBoostRegressor(random_state=0, verbose=100)
model = make_pipeline(
Profiler(),
Astype(),
NUniqueThreshold(max_freq=None),
DropCollinearFeatures(method="spearman", random_state=0),
ClippedFeatures(),
ModifiedStandardScaler(),
ModifiedSelectFromModel(model, random_state=0, threshold=1e-06),
CombinedFeatures(include_data=True),
ModifiedSelectFromModel(model, random_state=0, threshold=1e-06),
model,
)
model.fit(X, y)
Installation
pip install pretools
Testing
python setup.py test