Remove parameter normalize in linear models (#686)

CHANGELOG.md

@@ -37,7 +37,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 -
 -
 -
--
+- Remove parameter normalize in linear models ([#686](https://github.com/tinkoff-ai/etna/pull/686))
 -
 -
 

etna/models/linear.py

@@ -8,7 +8,7 @@
 class LinearPerSegmentModel(SklearnPerSegmentModel):
     """Class holding per segment :py:class:`sklearn.linear_model.LinearRegression`."""
 
-    def __init__(self, fit_intercept: bool = True, normalize: bool = False, **kwargs):
+    def __init__(self, fit_intercept: bool = True, **kwargs):
         """
         Create instance of LinearModel with given parameters.
 
@@ -17,25 +17,16 @@ def __init__(self, fit_intercept: bool = True, normalize: bool = False, **kwargs
         fit_intercept:
             Whether to calculate the intercept for this model. If set to False, no intercept will be used in
             calculations (i.e. data is expected to be centered).
-        normalize:
-            This parameter is ignored when ``fit_intercept`` is set to False.
-            If True, the regressors X will be normalized before regression
-            by subtracting the mean and dividing by the l2-norm.
         """
         self.fit_intercept = fit_intercept
-        self.normalize = normalize
         self.kwargs = kwargs
-        super().__init__(
-            regressor=LinearRegression(fit_intercept=self.fit_intercept, normalize=self.normalize, **self.kwargs)
-        )
+        super().__init__(regressor=LinearRegression(fit_intercept=self.fit_intercept, **self.kwargs))
 
 
 class ElasticPerSegmentModel(SklearnPerSegmentModel):
     """Class holding per segment :py:class:`sklearn.linear_model.ElasticNet`."""
 
-    def __init__(
-        self, alpha: float = 1.0, l1_ratio: float = 0.5, fit_intercept: bool = True, normalize: bool = False, **kwargs
-    ):
+    def __init__(self, alpha: float = 1.0, l1_ratio: float = 0.5, fit_intercept: bool = True, **kwargs):
         """
         Create instance of ElasticNet with given parameters.
 
@@ -58,21 +49,16 @@ def __init__(
         fit_intercept:
             Whether to calculate the intercept for this model. If set to False, no intercept will be used in
             calculations (i.e. data is expected to be centered).
-        normalize:
-            This parameter is ignored when fit_intercept is set to False. If True, the regressors X will be normalized
-            before regression by subtracting the mean and dividing by the l2-norm.
         """
         self.alpha = alpha
         self.l1_ratio = l1_ratio
         self.fit_intercept = fit_intercept
-        self.normalize = normalize
         self.kwargs = kwargs
         super().__init__(
             regressor=ElasticNet(
                 alpha=self.alpha,
                 l1_ratio=self.l1_ratio,
                 fit_intercept=self.fit_intercept,
-                normalize=self.normalize,
                 **self.kwargs,
             )
         )
@@ -81,7 +67,7 @@ def __init__(
 class LinearMultiSegmentModel(SklearnMultiSegmentModel):
     """Class holding :py:class:`sklearn.linear_model.LinearRegression` for all segments."""
 
-    def __init__(self, fit_intercept: bool = True, normalize: bool = False, **kwargs):
+    def __init__(self, fit_intercept: bool = True, **kwargs):
         """
         Create instance of LinearModel with given parameters.
 
@@ -90,25 +76,16 @@ def __init__(self, fit_intercept: bool = True, normalize: bool = False, **kwargs
         fit_intercept:
             Whether to calculate the intercept for this model. If set to False, no intercept will be used in
             calculations (i.e. data is expected to be centered).
-        normalize:
-            This parameter is ignored when ``fit_intercept`` is set to False.
-            If True, the regressors X will be normalized before regression
-            by subtracting the mean and dividing by the l2-norm.
         """
         self.fit_intercept = fit_intercept
-        self.normalize = normalize
         self.kwargs = kwargs
-        super().__init__(
-            regressor=LinearRegression(fit_intercept=self.fit_intercept, normalize=self.normalize, **self.kwargs)
-        )
+        super().__init__(regressor=LinearRegression(fit_intercept=self.fit_intercept, **self.kwargs))
 
 
 class ElasticMultiSegmentModel(SklearnMultiSegmentModel):
     """Class holding :py:class:`sklearn.linear_model.ElasticNet` for all segments."""
 
-    def __init__(
-        self, alpha: float = 1.0, l1_ratio: float = 0.5, fit_intercept: bool = True, normalize: bool = False, **kwargs
-    ):
+    def __init__(self, alpha: float = 1.0, l1_ratio: float = 0.5, fit_intercept: bool = True, **kwargs):
         """
         Create instance of ElasticNet with given parameters.
 
@@ -131,21 +108,16 @@ def __init__(
         fit_intercept:
             Whether to calculate the intercept for this model. If set to False, no intercept will be used in
             calculations (i.e. data is expected to be centered).
-        normalize:
-            This parameter is ignored when fit_intercept is set to False. If True, the regressors X will be normalized
-            before regression by subtracting the mean and dividing by the l2-norm.
         """
         self.alpha = alpha
         self.l1_ratio = l1_ratio
         self.fit_intercept = fit_intercept
-        self.normalize = normalize
         self.kwargs = kwargs
         super().__init__(
             regressor=ElasticNet(
                 alpha=self.alpha,
                 l1_ratio=self.l1_ratio,
                 fit_intercept=self.fit_intercept,
-                normalize=self.normalize,
                 **self.kwargs,
             )
         )

tests/test_models/test_linear_model.py

@@ -101,9 +101,9 @@ def test_repr_linear(model_class, model_class_repr):
     """Check __repr__ method of LinearPerSegmentModel and LinearMultiSegmentModel."""
     kwargs = {"copy_X": True, "positive": True}
     kwargs_repr = "copy_X = True, positive = True"
-    model = model_class(fit_intercept=True, normalize=False, **kwargs)
+    model = model_class(fit_intercept=True, **kwargs)
     model_repr = model.__repr__()
-    true_repr = f"{model_class_repr}(fit_intercept = True, normalize = False, {kwargs_repr}, )"
+    true_repr = f"{model_class_repr}(fit_intercept = True, {kwargs_repr}, )"
     assert model_repr == true_repr
 
 
@@ -115,11 +115,9 @@ def test_repr_elastic(model_class, model_class_repr):
     """Check __repr__ method of ElasticPerSegmentModel and ElasticMultiSegmentModel."""
     kwargs = {"copy_X": True, "positive": True}
     kwargs_repr = "copy_X = True, positive = True"
-    model = model_class(alpha=1.0, l1_ratio=0.5, fit_intercept=True, normalize=False, **kwargs)
+    model = model_class(alpha=1.0, l1_ratio=0.5, fit_intercept=True, **kwargs)
     model_repr = model.__repr__()
-    true_repr = (
-        f"{model_class_repr}(alpha = 1.0, l1_ratio = 0.5, " f"fit_intercept = True, normalize = False, {kwargs_repr}, )"
-    )
+    true_repr = f"{model_class_repr}(alpha = 1.0, l1_ratio = 0.5, " f"fit_intercept = True, {kwargs_repr}, )"
     assert model_repr == true_repr