[WIP] DOC - add example of scikit-learn GridSearchCV

examples/grid_search_cv.py

@@ -0,0 +1,35 @@
+"""
+==========================================
+Skglm support of scikit-learn GridSearchCV
+==========================================
+An example that uses scikit-learn``GridSearchCV`` to select 
+the best ``alpha`` and ``l1_ratio`` of ElasticNet model.
+"""
+
+import numpy as np
+from skglm import ElasticNet
+from skglm.utils import make_correlated_data
+from sklearn.model_selection import GridSearchCV
+
+
+# Simulate dataset
+n_samples, n_features = 10, 100
+X, y, _ = make_correlated_data(n_samples, n_features, random_state=0)
+
+# grid of parameter
+alpha_max = np.linalg.norm(X.T @ y, ord=np.inf) / n_samples
+parameters = {
+    'alpha': alpha_max * np.geomspace(1, 1e-3, 100),
+    'l1_ratio': [1., 0.9, 0.8, 0.7]
+}
+
+# init and fit GridSearchCV
+reg = GridSearchCV(
+    ElasticNet(),
+    param_grid=parameters,
+    cv=5, n_jobs=-1
+)
+reg.fit(X, y)
+
+# print the best parameters
+print(reg.best_estimator_)