05H.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Python ≥3.5 is required\n",
    "import sys\n",
    "assert sys.version_info >= (3, 5)\n",
    "\n",
    "# Scikit-Learn ≥0.20 is required\n",
    "import sklearn\n",
    "assert sklearn.__version__ >= \"0.20\"\n",
    "\n",
    "# Common imports\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "# to make this notebook's output stable across runs\n",
    "np.random.seed(42)\n",
    "\n",
    "# To plot pretty figures\n",
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "mpl.rc('axes', labelsize=14)\n",
    "mpl.rc('xtick', labelsize=12)\n",
    "mpl.rc('ytick', labelsize=12)\n",
    "\n",
    "# Where to save the figures\n",
    "PROJECT_ROOT_DIR = \".\"\n",
    "CHAPTER_ID = \"svm\"\n",
    "IMAGES_PATH = os.path.join(PROJECT_ROOT_DIR, \"images\", CHAPTER_ID)\n",
    "os.makedirs(IMAGES_PATH, exist_ok=True)\n",
    "\n",
    "def save_fig(fig_id, tight_layout=True, fig_extension=\"png\", resolution=300):\n",
    "    path = os.path.join(IMAGES_PATH, fig_id + \".\" + fig_extension)\n",
    "    print(\"Saving figure\", fig_id)\n",
    "    if tight_layout:\n",
    "        plt.tight_layout()\n",
    "    plt.savefig(path, format=fig_extension, dpi=resolution)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import datasets\n",
    "\n",
    "iris = datasets.load_iris()\n",
    "X = iris[\"data\"][:, (2, 3)]\n",
    "y = iris[\"target\"]\n",
    "\n",
    "setosa_or_versicolor = (y == 0) | (y == 1)\n",
    "X = X[setosa_or_versicolor]\n",
    "y = y[setosa_or_versicolor]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((100, 2), (100,))"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearSVC:                    [0.28474272] [[1.05364736 1.09903308]]\n",
      "SVC:                          [0.31896852] [[1.1203284  1.02625193]]\n",
      "SGDClassifier(alpha=0.00200): [0.117] [[0.77714169 0.72981762]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.svm import SVC, LinearSVC\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "C = 5\n",
    "alpha = 1 / (C * len(X))\n",
    "\n",
    "lin_clf = LinearSVC(loss=\"hinge\", C=C, random_state=42)\n",
    "svm_clf = SVC(kernel=\"linear\", C=C)\n",
    "sgd_clf = SGDClassifier(loss=\"hinge\", learning_rate=\"constant\",\n",
    "                        eta0=0.001, alpha=alpha, max_iter=1000,\n",
    "                        tol=1e-3, random_state=42)\n",
    "\n",
    "scaler = StandardScaler()\n",
    "X_scaled = scaler.fit_transform(X)\n",
    "\n",
    "lin_clf.fit(X_scaled, y)\n",
    "svm_clf.fit(X_scaled, y)\n",
    "sgd_clf.fit(X_scaled, y)\n",
    "\n",
    "print(\"LinearSVC:                   \", lin_clf.intercept_, lin_clf.coef_)\n",
    "print(\"SVC:                         \", svm_clf.intercept_, svm_clf.coef_)\n",
    "print(\"SGDClassifier(alpha={:.5f}):\".format(sgd_clf.alpha), sgd_clf.intercept_, sgd_clf.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Compute the slope and bias of each decision boundary\n",
    "w1 = -lin_clf.coef_[0, 0]/lin_clf.coef_[0, 1]\n",
    "b1 = -lin_clf.intercept_[0]/lin_clf.coef_[0, 1]\n",
    "w2 = -svm_clf.coef_[0, 0]/svm_clf.coef_[0, 1]\n",
    "b2 = -svm_clf.intercept_[0]/svm_clf.coef_[0, 1]\n",
    "w3 = -sgd_clf.coef_[0, 0]/sgd_clf.coef_[0, 1]\n",
    "b3 = -sgd_clf.intercept_[0]/sgd_clf.coef_[0, 1]\n",
    "\n",
    "# Transform the decision boundary lines back to the original scale\n",
    "line1 = scaler.inverse_transform([[-10, -10 * w1 + b1], [10, 10 * w1 + b1]])\n",
    "line2 = scaler.inverse_transform([[-10, -10 * w2 + b2], [10, 10 * w2 + b2]])\n",
    "line3 = scaler.inverse_transform([[-10, -10 * w3 + b3], [10, 10 * w3 + b3]])\n",
    "\n",
    "# Plot all three decision boundaries\n",
    "plt.figure(figsize=(11, 4))\n",
    "plt.plot(line1[:, 0], line1[:, 1], \"k:\", label=\"LinearSVC\")\n",
    "plt.plot(line2[:, 0], line2[:, 1], \"b--\", linewidth=2, label=\"SVC\")\n",
    "plt.plot(line3[:, 0], line3[:, 1], \"r-\", label=\"SGDClassifier\")\n",
    "plt.plot(X[:, 0][y==1], X[:, 1][y==1], \"bs\") # label=\"Iris versicolor\"\n",
    "plt.plot(X[:, 0][y==0], X[:, 1][y==0], \"yo\") # label=\"Iris setosa\"\n",
    "plt.xlabel(\"Petal length\", fontsize=14)\n",
    "plt.ylabel(\"Petal width\", fontsize=14)\n",
    "plt.legend(loc=\"upper center\", fontsize=14)\n",
    "plt.axis([0, 5.5, 0, 2])\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_openml\n",
    "\n",
    "mnist = fetch_openml(\"mnist_784\", version=1, cache=True)\n",
    "\n",
    "X = mnist[\"data\"]\n",
    "y = mnist[\"target\"].astype(np.uint8)\n",
    "\n",
    "X_train, y_train = X[:60000], y[:60000]\n",
    "X_test, y_test   = X[60000:], y[60000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/biowar1337/anaconda3/lib/python3.7/site-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "          intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
       "          multi_class='ovr', penalty='l2', random_state=42, tol=0.0001,\n",
       "          verbose=0)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lin_clf = LinearSVC(random_state=42)\n",
    "lin_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8733"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "y_pred = lin_clf.predict(X_train)\n",
    "accuracy_score(y_train, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "X_train_scaled = scaler.fit_transform(X_train.astype(np.float32))\n",
    "X_test_scaled = scaler.fit_transform(X_test.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/biowar1337/anaconda3/lib/python3.7/site-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "          intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
       "          multi_class='ovr', penalty='l2', random_state=42, tol=0.0001,\n",
       "          verbose=0)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lin_clf = LinearSVC(random_state=42)\n",
    "lin_clf.fit(X_train_scaled, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9226833333333333"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = lin_clf.predict(X_train_scaled)\n",
    "accuracy_score(y_train, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=1.0, break_ties=False, cache_size=200, class_weight=None, coef0=0.0,\n",
       "    decision_function_shape='ovr', degree=3, gamma='scale', kernel='rbf',\n",
       "    max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "    tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_clf = SVC(gamma=\"scale\")\n",
    "svm_clf.fit(X_train_scaled[:10000], y_train[:10000])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9455333333333333"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = svm_clf.predict(X_train_scaled)\n",
    "accuracy_score(y_train, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 10 candidates, totalling 30 fits\n",
      "[CV] C=10.50714306409916, gamma=0.029106359131330688 .................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .. C=10.50714306409916, gamma=0.029106359131330688, total=   1.6s\n",
      "[CV] C=10.50714306409916, gamma=0.029106359131330688 .................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    1.6s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .. C=10.50714306409916, gamma=0.029106359131330688, total=   1.6s\n",
      "[CV] C=10.50714306409916, gamma=0.029106359131330688 .................\n",
      "[CV] .. C=10.50714306409916, gamma=0.029106359131330688, total=   1.6s\n",
      "[CV] C=6.986584841970366, gamma=0.0020513382630874496 ................\n",
      "[CV] . C=6.986584841970366, gamma=0.0020513382630874496, total=   1.3s\n",
      "[CV] C=6.986584841970366, gamma=0.0020513382630874496 ................\n",
      "[CV] . C=6.986584841970366, gamma=0.0020513382630874496, total=   1.3s\n",
      "[CV] C=6.986584841970366, gamma=0.0020513382630874496 ................\n",
      "[CV] . C=6.986584841970366, gamma=0.0020513382630874496, total=   1.3s\n",
      "[CV] C=2.5599452033620267, gamma=0.0013066739238053278 ...............\n",
      "[CV]  C=2.5599452033620267, gamma=0.0013066739238053278, total=   1.2s\n",
      "[CV] C=2.5599452033620267, gamma=0.0013066739238053278 ...............\n",
      "[CV]  C=2.5599452033620267, gamma=0.0013066739238053278, total=   1.2s\n",
      "[CV] C=2.5599452033620267, gamma=0.0013066739238053278 ...............\n",
      "[CV]  C=2.5599452033620267, gamma=0.0013066739238053278, total=   1.2s\n",
      "[CV] C=9.661761457749352, gamma=0.01593052261624101 ..................\n",
      "[CV] ... C=9.661761457749352, gamma=0.01593052261624101, total=   1.6s\n",
      "[CV] C=9.661761457749352, gamma=0.01593052261624101 ..................\n",
      "[CV] ... C=9.661761457749352, gamma=0.01593052261624101, total=   1.8s\n",
      "[CV] C=9.661761457749352, gamma=0.01593052261624101 ..................\n",
      "[CV] ... C=9.661761457749352, gamma=0.01593052261624101, total=   1.7s\n",
      "[CV] C=8.080725777960454, gamma=0.0010994335574766197 ................\n",
      "[CV] . C=8.080725777960454, gamma=0.0010994335574766197, total=   1.2s\n",
      "[CV] C=8.080725777960454, gamma=0.0010994335574766197 ................\n",
      "[CV] . C=8.080725777960454, gamma=0.0010994335574766197, total=   1.1s\n",
      "[CV] C=8.080725777960454, gamma=0.0010994335574766197 ................\n",
      "[CV] . C=8.080725777960454, gamma=0.0010994335574766197, total=   1.2s\n",
      "[CV] C=10.699098521619943, gamma=0.046225890010208284 ................\n",
      "[CV] . C=10.699098521619943, gamma=0.046225890010208284, total=   1.8s\n",
      "[CV] C=10.699098521619943, gamma=0.046225890010208284 ................\n",
      "[CV] . C=10.699098521619943, gamma=0.046225890010208284, total=   1.8s\n",
      "[CV] C=10.699098521619943, gamma=0.046225890010208284 ................\n",
      "[CV] . C=10.699098521619943, gamma=0.046225890010208284, total=   1.8s\n",
      "[CV] C=3.1233911067827616, gamma=0.002310201887845293 ................\n",
      "[CV] . C=3.1233911067827616, gamma=0.002310201887845293, total=   1.5s\n",
      "[CV] C=3.1233911067827616, gamma=0.002310201887845293 ................\n",
      "[CV] . C=3.1233911067827616, gamma=0.002310201887845293, total=   1.5s\n",
      "[CV] C=3.1233911067827616, gamma=0.002310201887845293 ................\n",
      "[CV] . C=3.1233911067827616, gamma=0.002310201887845293, total=   1.4s\n",
      "[CV] C=2.834045098534338, gamma=0.004059611610484304 .................\n",
      "[CV] .. C=2.834045098534338, gamma=0.004059611610484304, total=   1.7s\n",
      "[CV] C=2.834045098534338, gamma=0.004059611610484304 .................\n",
      "[CV] .. C=2.834045098534338, gamma=0.004059611610484304, total=   1.6s\n",
      "[CV] C=2.834045098534338, gamma=0.004059611610484304 .................\n",
      "[CV] .. C=2.834045098534338, gamma=0.004059611610484304, total=   1.6s\n",
      "[CV] C=6.247564316322379, gamma=0.0073095398359129095 ................\n",
      "[CV] . C=6.247564316322379, gamma=0.0073095398359129095, total=   1.6s\n",
      "[CV] C=6.247564316322379, gamma=0.0073095398359129095 ................\n",
      "[CV] . C=6.247564316322379, gamma=0.0073095398359129095, total=   1.6s\n",
      "[CV] C=6.247564316322379, gamma=0.0073095398359129095 ................\n",
      "[CV] . C=6.247564316322379, gamma=0.0073095398359129095, total=   1.6s\n",
      "[CV] C=3.9122914019804194, gamma=0.016738085788752122 ................\n",
      "[CV] . C=3.9122914019804194, gamma=0.016738085788752122, total=   1.6s\n",
      "[CV] C=3.9122914019804194, gamma=0.016738085788752122 ................\n",
      "[CV] . C=3.9122914019804194, gamma=0.016738085788752122, total=   1.6s\n",
      "[CV] C=3.9122914019804194, gamma=0.016738085788752122 ................\n",
      "[CV] . C=3.9122914019804194, gamma=0.016738085788752122, total=   1.6s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done  30 out of  30 | elapsed:   45.2s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, error_score=nan,\n",
       "                   estimator=SVC(C=1.0, break_ties=False, cache_size=200,\n",
       "                                 class_weight=None, coef0=0.0,\n",
       "                                 decision_function_shape='ovr', degree=3,\n",
       "                                 gamma='scale', kernel='rbf', max_iter=-1,\n",
       "                                 probability=False, random_state=None,\n",
       "                                 shrinking=True, tol=0.001, verbose=False),\n",
       "                   iid='deprecated', n_iter=10, n_jobs=None,\n",
       "                   param_distributions={'C': <scipy.stats._distn_infrastructure.rv_frozen object at 0x7f8f44fc9850>,\n",
       "                                        'gamma': <scipy.stats._distn_infrastructure.rv_frozen object at 0x7f8f40cafb90>},\n",
       "                   pre_dispatch='2*n_jobs', random_state=None, refit=True,\n",
       "                   return_train_score=False, scoring=None, verbose=2)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import reciprocal, uniform\n",
    "\n",
    "param_distributions = {\"gamma\": reciprocal(0.001, 0.1),\n",
    "                       \"C\": uniform(1, 10)}\n",
    "rnd_search_cv = RandomizedSearchCV(svm_clf, param_distributions,\n",
    "                                   n_iter=10, verbose=2, cv=3)\n",
    "rnd_search_cv.fit(X_train_scaled[:1000], y_train[:1000])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=8.080725777960454, break_ties=False, cache_size=200, class_weight=None,\n",
       "    coef0=0.0, decision_function_shape='ovr', degree=3,\n",
       "    gamma=0.0010994335574766197, kernel='rbf', max_iter=-1, probability=False,\n",
       "    random_state=None, shrinking=True, tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_search_cv.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8619907332482183"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_search_cv.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "#rnd_search_cv.best_estimator_.fit(X_train_scaled, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}