UserWarning: X has feature names, but PowerTransformer was fitted without feature names #31947
Description
Describe the bug
When using pandas dataframes and a TransformedTargetRegressor with PowerTransformer with set_output(transform="pandas"), I get this warning:
UserWarning: X has feature names, but PowerTransformer was fitted without feature names
The warning does not arise when using other estimators (e.g. StandardScaler) but only with PowerTransformer.
The problem seems to originate from the inverse_transform implementation of PowerTransformer.
Steps/Code to Reproduce
from sklearn.datasets import load_iris
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.compose import TransformedTargetRegressor
from sklearn.preprocessing import PowerTransformer, StandardScaler
X, y = load_iris(return_X_y=True, as_frame=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)
# This works fine:
pipeline = TransformedTargetRegressor(
regressor=LinearRegression(),
transformer=StandardScaler().set_output(transform="pandas")
)
pipeline.fit(X_train, y_train)
y_test_pred = pipeline.predict(X_test)
# But this gets a warning:
pipeline = TransformedTargetRegressor(
regressor=LinearRegression(),
transformer=PowerTransformer().set_output(transform="pandas")
)
pipeline.fit(X_train, y_train)
y_test_pred = pipeline.predict(X_test)Expected Results
No warning
Actual Results
UserWarning: X has feature names, but PowerTransformer was fitted without feature names
Versions
System:
python: 3.11.9
Python dependencies:
sklearn: 1.7.1
pip: 25.2
setuptools: 65.5.0
numpy: 2.0.2
scipy: 1.15.1
Cython: None
pandas: 2.2.3
matplotlib: 3.10.0
joblib: 1.4.2
threadpoolctl: 3.5.0