DOC ensures MultiTaskLasso passes numpydoc validation (scikit-learn#21363)

felixglush · glemaitre · commit a734af1a7662 · 2021-10-23T12:02:30.000+02:00
diff --git a/maint_tools/test_docstrings.py b/maint_tools/test_docstrings.py
@@ -16,7 +16,6 @@
     "MultiLabelBinarizer",
     "MultiTaskElasticNet",
     "MultiTaskElasticNetCV",
-    "MultiTaskLasso",
     "MultiTaskLassoCV",
     "OrthogonalMatchingPursuit",
     "OrthogonalMatchingPursuitCV",
diff --git a/sklearn/linear_model/_coordinate_descent.py b/sklearn/linear_model/_coordinate_descent.py
@@ -2378,6 +2378,7 @@ def fit(self, X, y):
         Returns
         -------
         self : object
+            Fitted estimator.
 
         Notes
         -----
@@ -2527,7 +2528,7 @@ class MultiTaskLasso(MultiTaskElasticNet):
         If set to 'random', a random coefficient is updated every iteration
         rather than looping over features sequentially by default. This
         (setting to 'random') often leads to significantly faster convergence
-        especially when tol is higher than 1e-4
+        especially when tol is higher than 1e-4.
 
     Attributes
     ----------
@@ -2563,6 +2564,19 @@ class MultiTaskLasso(MultiTaskElasticNet):
 
         .. versionadded:: 1.0
 
+    See Also
+    --------
+    Lasso: Linear Model trained with L1 prior as regularizer (aka the Lasso).
+    MultiTaskLasso: Multi-task L1/L2 Lasso with built-in cross-validation.
+    MultiTaskElasticNet: Multi-task L1/L2 ElasticNet with built-in cross-validation.
+
+    Notes
+    -----
+    The algorithm used to fit the model is coordinate descent.
+
+    To avoid unnecessary memory duplication the X and y arguments of the fit
+    method should be directly passed as Fortran-contiguous numpy arrays.
+
     Examples
     --------
     >>> from sklearn import linear_model
@@ -2574,19 +2588,6 @@ class MultiTaskLasso(MultiTaskElasticNet):
     [0.         0.94592424]]
     >>> print(clf.intercept_)
     [-0.41888636 -0.87382323]
-
-    See Also
-    --------
-    MultiTaskLasso : Multi-task L1/L2 Lasso with built-in cross-validation
-    Lasso
-    MultiTaskElasticNet
-
-    Notes
-    -----
-    The algorithm used to fit the model is coordinate descent.
-
-    To avoid unnecessary memory duplication the X and y arguments of the fit
-    method should be directly passed as Fortran-contiguous numpy arrays.
     """
 
     def __init__(
