SklearnDataset¶
Wrapper for scikit-learn compatible data providing factrainer-compatible interface.
This class wraps feature matrix and target vector to provide a consistent
interface for cross-validation and data manipulation within the factrainer
framework. It supports multiple data formats including numpy arrays, pandas
DataFrames/Series, and polars DataFrames/Series.
This dataset is passed to the train and predict methods of
SingleModelContainer and CvModelContainer when using scikit-learn models.
Attributes:
| Name | Type | Description |
|---|---|---|
X |
MatrixLike
|
Feature matrix. Can be a numpy array, pandas DataFrame, or polars DataFrame. |
y |
VectorLike | None
|
Target vector. Can be a numpy array, pandas Series, polars Series, or None. |
Examples:
>>> import numpy as np
>>> from factrainer.sklearn import SklearnDataset
>>> # Create from numpy arrays
>>> X = np.random.randn(100, 10)
>>> y = np.random.randn(100)
>>> dataset = SklearnDataset(X=X, y=y)
>>> # Create without target
>>> dataset = SklearnDataset(X=X)
>>> # Create from pandas
>>> import pandas as pd
>>> df = pd.DataFrame(X, columns=[f"feature_{i}" for i in range(10)])
>>> target = pd.Series(y, name="target")
>>> dataset = SklearnDataset(X=df, y=target)