Remove unnecessary Custom classes

stratestic/strategies/machine_learning/helpers/__init__.py

@@ -8,6 +8,4 @@
 from stratestic.strategies.machine_learning.helpers._pipeline_custom_classes import (
     FeatureSelector,
     CustomOneHotEncoder,
-    CustomStandardScaler,
-    CustomPolynomialFeatures
 )

stratestic/strategies/machine_learning/helpers/_pipeline_custom_classes.py

@@ -1,8 +1,7 @@
 import pandas as pd
 
 from sklearn.base import BaseEstimator, TransformerMixin
-from sklearn.pipeline import Pipeline
-from sklearn.preprocessing import OneHotEncoder, StandardScaler, PolynomialFeatures
+from sklearn.preprocessing import OneHotEncoder
 
 
 class FeatureSelector(BaseEstimator, TransformerMixin):
@@ -90,69 +89,3 @@ def transform(self, X, y=None, **transform_params):
         transformed_data = super(CustomOneHotEncoder, self).transform(X, **transform_params).toarray()
 
         return pd.DataFrame(data=transformed_data, columns=self.columns)
-
-
-class CustomStandardScaler(StandardScaler):
-    """
-    A custom implementation of StandardScaler to include feature name handling.
-
-    This class extends the StandardScaler to retain feature names for downstream use in the pipeline.
-
-    Parameters
-    ----------
-    None
-
-    Attributes
-    ----------
-    features : list of str
-        Feature names after scaling.
-
-    Methods
-    -------
-    fit(X, y=None, **fit_params)
-        Fits the scaler to X and stores feature names.
-    transform(X, y=None, **transform_params)
-        Scales features of X using the mean and standard deviation.
-    get_feature_names()
-        Returns the feature names after transformation.
-    """
-
-    def fit(self, X, y=None, **fit_params):
-        self.features = X.columns
-
-        return super(CustomStandardScaler, self).fit(X, y, **fit_params)
-
-
-class CustomPolynomialFeatures(PolynomialFeatures):
-    """
-    A custom implementation of PolynomialFeatures that retains original feature names.
-
-    This class extends PolynomialFeatures to include original feature names in the output DataFrame, facilitating interpretability.
-
-    Parameters
-    ----------
-    None
-
-    Attributes
-    ----------
-    columns : list of str
-        Original feature names.
-
-    Methods
-    -------
-    fit(X, y=None)
-        Fits the transformer to X and stores original feature names.
-    transform(X, **transform_params)
-        Transforms X by adding polynomial features and returns a DataFrame with appropriate feature names.
-    """
-
-    def fit(self, X, y=None, **kwargs):
-        self.columns = X.columns
-
-        return super(CustomPolynomialFeatures, self).fit(X, y)
-
-    def transform(self, X, **transform_params):
-
-        transformed_data = super(CustomPolynomialFeatures, self).transform(X)
-
-        return pd.DataFrame(data=transformed_data, columns=self.columns)

stratestic/strategies/machine_learning/helpers/_training.py

@@ -5,6 +5,7 @@
 from sklearn.base import is_classifier
 from sklearn.model_selection import GridSearchCV
 from sklearn.pipeline import Pipeline, FeatureUnion
+from sklearn.preprocessing import StandardScaler, PolynomialFeatures
 
 from stratestic.strategies.machine_learning.helpers._defaults import estimator_mapping, estimator_params
 from stratestic.strategies.machine_learning.helpers._helpers import train_test_split_ts
@@ -13,8 +14,6 @@
 from stratestic.strategies.machine_learning.helpers._pipeline_custom_classes import (
     FeatureSelector,
     CustomOneHotEncoder,
-    CustomStandardScaler,
-    CustomPolynomialFeatures
 )
 
 grid_search_params_defaults = {
@@ -27,7 +26,7 @@
 configure_logger()
 
 
-def build_pipeline(estimator, polynomial_degree=1):
+def build_pipeline(estimator, polynomial_degree=1, interaction_only=False):
     """
     Constructs a machine learning pipeline with optional grid search capability.
 
@@ -37,6 +36,10 @@ def build_pipeline(estimator, polynomial_degree=1):
         The estimator (machine learning algorithm) to use.
     polynomial_degree : int, optional
         Degree of polynomial features to generate. Default is 1.
+    interaction_only : bool, default=False
+        If `True`, only interaction features are produced: features that are
+        products of at most `degree` *distinct* input features, i.e. terms with
+        power of 2 or higher of the same input feature are excluded.
 
     Returns
     -------
@@ -52,8 +55,12 @@ def build_pipeline(estimator, polynomial_degree=1):
         ('features', FeatureUnion([
             ('num_features', Pipeline([
                 ('selector', FeatureSelector('num')),
-                ('feature_polynomials', CustomPolynomialFeatures(degree=polynomial_degree, include_bias=False)),
-                ('scaling', CustomStandardScaler())
+                ('feature_polynomials', PolynomialFeatures(
+                    degree=polynomial_degree,
+                    include_bias=False,
+                    interaction_only=interaction_only
+                )),
+                ('scaling', StandardScaler())
             ])),
             ('cat_features', Pipeline([
                 ('selector', FeatureSelector('cat')),