API : set return_models to False by default in path functions in coordinate descent

Author: agramfort

examples/linear_model/plot_lasso_coordinate_descent_path.py

@@ -30,24 +30,18 @@
 eps = 5e-3  # the smaller it is the longer is the path
 
 print("Computing regularization path using the lasso...")
-# The return_models parameter sets that lasso_path will return
-# the alphas and the coefficients as output, instead of a list
-# of models as it does by default. Returning the list of models
-# is deprecated and will eventually be removed in 0.15
-alphas_lasso, coefs_lasso, _ = lasso_path(X, y, eps, return_models=False,
-                                          fit_intercept=False)
+alphas_lasso, coefs_lasso, _ = lasso_path(X, y, eps, fit_intercept=False)
 
 print("Computing regularization path using the positive lasso...")
 alphas_positive_lasso, coefs_positive_lasso, _ = lasso_path(
-    X, y, eps, positive=True, return_models=False, fit_intercept=False)
+    X, y, eps, positive=True, fit_intercept=False)
 print("Computing regularization path using the elastic net...")
 alphas_enet, coefs_enet, _ = enet_path(
-    X, y, eps=eps, l1_ratio=0.8, return_models=False, fit_intercept=False)
+    X, y, eps=eps, l1_ratio=0.8, fit_intercept=False)
 
 print("Computing regularization path using the positve elastic net...")
 alphas_positive_enet, coefs_positive_enet, _ = enet_path(
-    X, y, eps=eps, l1_ratio=0.8, positive=True, return_models=False,
-    fit_intercept=False)
+    X, y, eps=eps, l1_ratio=0.8, positive=True, fit_intercept=False)
 
 # Display results
 

sklearn/linear_model/coordinate_descent.py

@@ -90,7 +90,7 @@ def _alpha_grid(X, y, Xy=None, l1_ratio=1.0, fit_intercept=True,
 def lasso_path(X, y, eps=1e-3, n_alphas=100, alphas=None,
                precompute='auto', Xy=None, fit_intercept=None,
                normalize=None, copy_X=True, coef_init=None,
-               verbose=False, return_models=True,
+               verbose=False, return_models=False,
                **params):
     """Compute Lasso path with coordinate descent
 
@@ -200,7 +200,7 @@ def lasso_path(X, y, eps=1e-3, n_alphas=100, alphas=None,
     >>> y = np.array([1, 2, 3.1])
     >>> # Use lasso_path to compute a coefficient path
     >>> _, coef_path, _ = lasso_path(X, y, alphas=[5., 1., .5],
-    ...                              return_models=False, fit_intercept=False)
+    ...                              fit_intercept=False)
     >>> print(coef_path)
     [[ 0.          0.          0.46874778]
      [ 0.2159048   0.4425765   0.23689075]]
@@ -236,7 +236,7 @@ def lasso_path(X, y, eps=1e-3, n_alphas=100, alphas=None,
 def enet_path(X, y, l1_ratio=0.5, eps=1e-3, n_alphas=100, alphas=None,
               precompute='auto', Xy=None, fit_intercept=True,
               normalize=False, copy_X=True, coef_init=None,
-              verbose=False, return_models=True,
+              verbose=False, return_models=False,
               **params):
     """Compute Elastic-Net path with coordinate descent
 
@@ -296,7 +296,7 @@ def enet_path(X, y, l1_ratio=0.5, eps=1e-3, n_alphas=100, alphas=None,
     verbose : bool or integer
         Amount of verbosity
 
-    return_models : boolean, optional, default True
+    return_models : boolean, optional, default False
         If ``True``, the function will return list of models. Setting it
         to ``False`` will change the function output returning the values
         of the alphas and the coefficients along the path. Returning the
@@ -628,7 +628,7 @@ def fit(self, X, y):
                           precompute=precompute, Xy=this_Xy,
                           fit_intercept=False, normalize=False, copy_X=True,
                           verbose=False, tol=self.tol, positive=self.positive,
-                          return_models=False, X_mean=X_mean, X_std=X_std,
+                          X_mean=X_mean, X_std=X_std,
                           coef_init=coef_[k], max_iter=self.max_iter)
             coef_[k] = this_coef[:, 0]
             dual_gaps_[k] = this_dual_gap[0]
@@ -827,7 +827,6 @@ def _path_residuals(X, y, train, test, path, path_params, l1_ratio=1,
 
     # del path_params['precompute']
     path_params = path_params.copy()
-    path_params['return_models'] = False
     path_params['fit_intercept'] = False
     path_params['normalize'] = False
     path_params['Xy'] = Xy

sklearn/linear_model/tests/test_coordinate_descent.py

@@ -189,6 +189,7 @@ def test_lasso_path_return_models_vs_new_return_gives_same_coefficients():
     with warnings.catch_warnings():
         warnings.simplefilter("ignore", DeprecationWarning)
         coef_path = [e.coef_ for e in lasso_path(X, y, alphas=alphas,
+                                                 return_models=True,
                                                  fit_intercept=False)]
 
     # Use lars_path and lasso_path(new output) with 1D linear interpolation

sklearn/linear_model/tests/test_least_angle.py

@@ -331,15 +331,17 @@ def test_lasso_lars_vs_lasso_cd_ill_conditioned():
 
     _, lasso_coef2, _ = linear_model.lasso_path(X, y,
                                                 alphas=lars_alphas, tol=1e-6,
-                                                return_models=False,
                                                 fit_intercept=False)
 
     lasso_coef = np.zeros((w.shape[0], len(lars_alphas)))
-    for i, model in enumerate(linear_model.lasso_path(X, y,
-                                                      alphas=lars_alphas,
-                                                      tol=1e-6,
-                                                      fit_intercept=False)):
-        lasso_coef[:, i] = model.coef_
+    with warnings.catch_warnings():
+        warnings.simplefilter("ignore", DeprecationWarning)
+        for i, model in enumerate(linear_model.lasso_path(X, y,
+                                                          alphas=lars_alphas,
+                                                          tol=1e-6,
+                                                          return_models=True,
+                                                          fit_intercept=False)):
+            lasso_coef[:, i] = model.coef_
 
     np.testing.assert_array_almost_equal(lars_coef, lasso_coef, decimal=1)
     np.testing.assert_array_almost_equal(lars_coef, lasso_coef2, decimal=1)