From a5baa85363a048c8075b40fe3d1bc96fbdb855bc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Aur=C3=A9lien=20Geron?= <ageron@users.noreply.github.com>
Date: Tue, 14 Nov 2023 21:34:06 +1300
Subject: [PATCH] Set dual=True in LinearSVC and LinearSVR to avoid warning

---
 05_support_vector_machines.ipynb | 31 +++++++++++++++++++------------
 1 file changed, 19 insertions(+), 12 deletions(-)

diff --git a/05_support_vector_machines.ipynb b/05_support_vector_machines.ipynb
index 421a7d2..31cd016 100644
--- a/05_support_vector_machines.ipynb
+++ b/05_support_vector_machines.ipynb
@@ -358,6 +358,13 @@
     "plt.show()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Note: the default value for the `dual` hyperparameter of the `LinearSVC` and `LinearSVR` estimators will change from `True` to `\"auto\"` in Scikit-Learn 1.4, so I set `dual=True` throughout this notebook to ensure the output of this notebook remains unchanged."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 8,
@@ -387,7 +394,7 @@
     "y = (iris.target == 2)  # Iris virginica\n",
     "\n",
     "svm_clf = make_pipeline(StandardScaler(),\n",
-    "                        LinearSVC(C=1, random_state=42))\n",
+    "                        LinearSVC(C=1, dual=True, random_state=42))\n",
     "svm_clf.fit(X, y)"
    ]
   },
@@ -454,8 +461,8 @@
     "# extra code – this cell generates and saves Figure 5–4\n",
     "\n",
     "scaler = StandardScaler()\n",
-    "svm_clf1 = LinearSVC(C=1, max_iter=10_000, random_state=42)\n",
-    "svm_clf2 = LinearSVC(C=100, max_iter=10_000, random_state=42)\n",
+    "svm_clf1 = LinearSVC(C=1, max_iter=10_000, dual=True, random_state=42)\n",
+    "svm_clf2 = LinearSVC(C=100, max_iter=10_000, dual=True, random_state=42)\n",
     "\n",
     "scaled_svm_clf1 = make_pipeline(scaler, svm_clf1)\n",
     "scaled_svm_clf2 = make_pipeline(scaler, svm_clf2)\n",
@@ -595,7 +602,7 @@
     "polynomial_svm_clf = make_pipeline(\n",
     "    PolynomialFeatures(degree=3),\n",
     "    StandardScaler(),\n",
-    "    LinearSVC(C=10, max_iter=10_000, random_state=42)\n",
+    "    LinearSVC(C=10, max_iter=10_000, dual=True, random_state=42)\n",
     ")\n",
     "polynomial_svm_clf.fit(X, y)"
    ]
@@ -932,7 +939,7 @@
     "y = 4 + 3 * X[:, 0] + np.random.randn(50)\n",
     "\n",
     "svm_reg = make_pipeline(StandardScaler(),\n",
-    "                        LinearSVR(epsilon=0.5, random_state=42))\n",
+    "                        LinearSVR(epsilon=0.5, dual=True, random_state=42))\n",
     "svm_reg.fit(X, y)"
    ]
   },
@@ -978,7 +985,7 @@
     "    plt.axis(axes)\n",
     "\n",
     "svm_reg2 = make_pipeline(StandardScaler(),\n",
-    "                         LinearSVR(epsilon=1.2, random_state=42))\n",
+    "                         LinearSVR(epsilon=1.2, dual=True, random_state=42))\n",
     "svm_reg2.fit(X, y)\n",
     "\n",
     "svm_reg._support = find_support_vectors(svm_reg, X, y)\n",
@@ -1566,7 +1573,7 @@
     "scaler = StandardScaler()\n",
     "X_scaled = scaler.fit_transform(X)\n",
     "\n",
-    "lin_clf = LinearSVC(loss=\"hinge\", C=C, random_state=42).fit(X_scaled, y)\n",
+    "lin_clf = LinearSVC(loss=\"hinge\", C=C, dual=True, random_state=42).fit(X_scaled, y)\n",
     "svc_clf = SVC(kernel=\"linear\", C=C).fit(X_scaled, y)\n",
     "sgd_clf = SGDClassifier(alpha=alpha, random_state=42).fit(X_scaled, y)"
    ]
@@ -1975,7 +1982,7 @@
     }
    ],
    "source": [
-    "lin_clf = LinearSVC(random_state=42)\n",
+    "lin_clf = LinearSVC(dual=True, random_state=42)\n",
     "lin_clf.fit(X_train, y_train)"
    ]
   },
@@ -2011,7 +2018,7 @@
     }
    ],
    "source": [
-    "lin_clf = LinearSVC(max_iter=1_000_000, random_state=42)\n",
+    "lin_clf = LinearSVC(max_iter=1_000_000, dual=True, random_state=42)\n",
     "lin_clf.fit(X_train, y_train)"
    ]
   },
@@ -2090,7 +2097,7 @@
    ],
    "source": [
     "lin_clf = make_pipeline(StandardScaler(),\n",
-    "                        LinearSVC(random_state=42))\n",
+    "                        LinearSVC(dual=True, random_state=42))\n",
     "lin_clf.fit(X_train, y_train)"
    ]
   },
@@ -2347,7 +2354,7 @@
    "source": [
     "from sklearn.svm import LinearSVR\n",
     "\n",
-    "lin_svr = make_pipeline(StandardScaler(), LinearSVR(random_state=42))\n",
+    "lin_svr = make_pipeline(StandardScaler(), LinearSVR(dual=True, random_state=42))\n",
     "lin_svr.fit(X_train, y_train)"
    ]
   },
@@ -2377,7 +2384,7 @@
    ],
    "source": [
     "lin_svr = make_pipeline(StandardScaler(),\n",
-    "                        LinearSVR(max_iter=5000, random_state=42))\n",
+    "                        LinearSVR(max_iter=5000, dual=True, random_state=42))\n",
     "lin_svr.fit(X_train, y_train)"
    ]
   },