explicitly use liblinear solver for l1 penalty in logistic regression

Author: amueller

02-supervised-learning.ipynb

@@ -1309,7 +1309,7 @@
    ],
    "source": [
     "for C, marker in zip([0.001, 1, 100], ['o', '^', 'v']):\n",
-    "    lr_l1 = LogisticRegression(C=C, penalty=\"l1\").fit(X_train, y_train)\n",
+    "    lr_l1 = LogisticRegression(C=C, solver='liblinear', penalty=\"l1\").fit(X_train, y_train)\n",
     "    print(\"Training accuracy of l1 logreg with C={:.3f}: {:.2f}\".format(\n",
     "          C, lr_l1.score(X_train, y_train)))\n",
     "    print(\"Test accuracy of l1 logreg with C={:.3f}: {:.2f}\".format(\n",