Library to easily run Null Importances.
Null Importances is feature selection process using target permutation tests actual importance significance against the distribution of feature importances when fitted to noise (shuffled target).
from nullpom import run_null_importance
result = run_null_importance(
{"objective": "binary", "seed": 42},
X_train=X_train,
X_valid=X_valid,
y_train=y_train,
y_valid=y_valid,
)
pip install nullpom