From a57cac980cd4f155b4c80841b8c2c65cab43cd32 Mon Sep 17 00:00:00 2001
From: ChunkeySun <112787606+ChunkeySun@users.noreply.github.com>
Date: Thu, 27 Oct 2022 17:23:35 +0800
Subject: [PATCH] Update Chapter 4.ipynb

---
 Notebooks/Chapter 4.ipynb | 32 ++++++++++++++------------------
 1 file changed, 14 insertions(+), 18 deletions(-)

diff --git a/Notebooks/Chapter 4.ipynb b/Notebooks/Chapter 4.ipynb
index 3432898..5f633cf 100644
--- a/Notebooks/Chapter 4.ipynb	
+++ b/Notebooks/Chapter 4.ipynb	
@@ -178,11 +178,11 @@
     "df_no = df[df.default2 == 0].sample(frac=0.15)\n",
     "# Take all samples  where target value is 'yes'\n",
     "df_yes = df[df.default2 == 1]\n",
-    "df_ = df_no.append(df_yes)\n",
+    "df_ = df.concat([df_yes, df_no])\n",
     "\n",
     "ax1.scatter(df_[df_.default == 'Yes'].balance, df_[df_.default == 'Yes'].income, s=40, c='orange', marker='+',\n",
     "            linewidths=1)\n",
-    "ax1.scatter(df_[df_.default == 'No'].balance, df_[df_.default == 'No'].income, s=40, marker='o', linewidths='1',\n",
+    "ax1.scatter(df_[df_.default == 'No'].balance, df_[df_.default == 'No'].income, s=40, marker='o', linewidths=1,\n",
     "            edgecolors='lightblue', facecolors='white', alpha=.6)\n",
     "\n",
     "ax1.set_ylim(ymin=0)\n",
@@ -228,7 +228,7 @@
     "# and predicted classification.\n",
     "X_test = np.arange(df.balance.min(), df.balance.max()).reshape(-1,1)\n",
     "\n",
-    "clf = skl_lm.LogisticRegression(solver='newton-cg')\n",
+    "clf = skl_lm.LogisticRegression(solver='lbfgs')\n",
     "clf.fit(X_train,y)\n",
     "prob = clf.predict_proba(X_test)\n",
     "\n",
@@ -297,7 +297,7 @@
    "source": [
     "# Using newton-cg solver, the coefficients are equal/closest to the ones in the book. \n",
     "# I do not know the details on the differences between the solvers.\n",
-    "clf = skl_lm.LogisticRegression(solver='newton-cg')\n",
+    "clf = skl_lm.LogisticRegression(solver='lbfgs')\n",
     "X_train = df.balance.values.reshape(-1,1)\n",
     "clf.fit(X_train,y)\n",
     "print(clf)\n",
@@ -391,8 +391,7 @@
     }
    ],
    "source": [
-    "X_train = sm.add_constant(df.balance)\n",
-    "est = smf.Logit(y.ravel(), X_train).fit()\n",
+    "est = smf.logit(formula='default2 ~ 1 + balance', data=df).fit()\n",
     "est.summary2().tables[1]"
    ]
   },
@@ -481,10 +480,8 @@
     }
    ],
    "source": [
-    "X_train = sm.add_constant(df.student2)\n",
-    "y = df.default2\n",
     "\n",
-    "est = smf.Logit(y, X_train).fit()\n",
+    "est = smf.logit(formula='default2 ~ 1 + student2', data=df).fit()\n",
     "est.summary2().tables[1]"
    ]
   },
@@ -593,8 +590,7 @@
     }
    ],
    "source": [
-    "X_train = sm.add_constant(df[['balance', 'income', 'student2']])\n",
-    "est = smf.Logit(y, X_train).fit()\n",
+    "est = smf.logit(formula='default2 ~ 1 + student2', data=df).fit()\n",
     "est.summary2().tables[1]"
    ]
   },
@@ -622,8 +618,8 @@
     "# Vector with balance values for plotting\n",
     "X_test = np.arange(df.balance.min(), df.balance.max()).reshape(-1,1)\n",
     "\n",
-    "clf = skl_lm.LogisticRegression(solver='newton-cg')\n",
-    "clf2 = skl_lm.LogisticRegression(solver='newton-cg')\n",
+    "clf = skl_lm.LogisticRegression(solver='lbfgs')\n",
+    "clf2 = skl_lm.LogisticRegression(solver='lbfgs')\n",
     "\n",
     "clf.fit(X_train,y)\n",
     "clf2.fit(X_train2,y2)\n",
@@ -734,7 +730,7 @@
     "ax1.legend(loc=2)\n",
     "\n",
     "# Right plot\n",
-    "sns.boxplot('student', 'balance', data=df, orient='v', ax=ax2,  palette=c_palette);"
+    "sns.boxplot(x='student', y='balance', data=df, orient='v', ax=ax2,  palette=c_palette);"
    ]
   },
   {
@@ -808,8 +804,8 @@
     }
    ],
    "source": [
-    "X = df[['balance', 'income', 'student2']].as_matrix()\n",
-    "y = df.default2.as_matrix()\n",
+    "X = df[['balance', 'income', 'student2']].values\n",
+    "y = df.default2.values\n",
     "\n",
     "lda = LinearDiscriminantAnalysis(solver='svd')\n",
     "y_pred = lda.fit(X, y).predict(X)\n",
@@ -1310,8 +1306,8 @@
     "X_test = X_scaled[:1000,:]\n",
     "y_test = y[:1000]\n",
     "\n",
-    "def KNN(n_neighbors=1, weights='uniform'):\n",
-    "    clf = neighbors.KNeighborsClassifier(n_neighbors, weights)\n",
+    "def KNN(n_neighbors=1,):\n",
+    "    clf = neighbors.KNeighborsClassifier(n_neighbors, weights = 'uniform')\n",
     "    clf.fit(X_train, y_train)\n",
     "    pred = clf.predict(X_test)\n",
     "    score = clf.score(X_test, y_test)\n",