488 lines
123 KiB
Text
488 lines
123 KiB
Text
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"id": "3eb7a65b",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"####### Import #######\n",
|
|
"from sklearn.datasets import fetch_openml\n",
|
|
"import sklearn\n",
|
|
"from matplotlib import pyplot as plt\n",
|
|
"from sklearn import model_selection\n",
|
|
"from sklearn import neural_network\n",
|
|
"from sklearn import metrics\n",
|
|
"from sklearn.svm import SVC\n",
|
|
"import numpy as np\n",
|
|
"import time\n",
|
|
"import warnings\n",
|
|
"warnings.filterwarnings(\"ignore\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 3,
|
|
"id": "a8812842",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"####### Loading data #######\n",
|
|
"mnist = fetch_openml('mnist_784',as_frame=False)\n",
|
|
"# images = mnist.data.reshape((-1, 28, 28))\n",
|
|
"# plt.imshow(images[0],cmap=plt.cm.gray_r,interpolation=\"nearest\")\n",
|
|
"# plt.show()\n",
|
|
"# print(\"Classe : \", mnist.target[0])"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 4,
|
|
"id": "6ec263be",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Dataset size : 5000\n",
|
|
"Etiquettes size : 5000\n",
|
|
"xtrain size : 4500\n",
|
|
"xtest size : 500\n",
|
|
"ytrain size : 4500\n",
|
|
"ytest size : 500\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"### Create vector of 1000 random indexes\n",
|
|
"rand_indexes = np.random.randint(70000, size=5000)\n",
|
|
"### Load data with the previous vector\n",
|
|
"data = mnist.data[rand_indexes]\n",
|
|
"print(\"Dataset size : \", len(data))\n",
|
|
"target = mnist.target[rand_indexes]\n",
|
|
"print(\"Etiquettes size : \", len(target))\n",
|
|
"\n",
|
|
"### Split the dataset for training and testing\n",
|
|
"# xtrain data set d'entraînement et ytrain étiquettes de xtrain\n",
|
|
"# xtest dataset de prédiction et ytest étiquettes de xtest\n",
|
|
"xtrain, xtest, ytrain, ytest = model_selection.train_test_split(data, target,train_size=0.9)\n",
|
|
"print(\"xtrain size : \", len(xtrain))\n",
|
|
"print(\"xtest size : \", len(xtest))\n",
|
|
"print(\"ytrain size : \", len(ytrain))\n",
|
|
"print(\"ytest size : \", len(ytest))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 7,
|
|
"id": "3b1a54ef",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Temps d'entraînement : 0.28424\n",
|
|
"Score échantillon de test : 0.906\n",
|
|
"Classe image 4 : 7\n",
|
|
"Classe prédite image 4 : 9\n",
|
|
"Précision pour chaque classe : \n",
|
|
" [0.96078431 0.88235294 0.89830508 0.89361702 0.94117647 0.87179487\n",
|
|
" 0.92592593 0.90243902 0.9 0.875 ]\n",
|
|
"Matrice de confusion SVM:\n",
|
|
" [[49 0 0 0 0 0 1 0 0 0]\n",
|
|
" [ 0 60 0 0 0 0 0 0 0 1]\n",
|
|
" [ 1 0 53 0 0 0 0 0 0 0]\n",
|
|
" [ 0 1 2 42 0 2 0 0 1 0]\n",
|
|
" [ 0 0 1 0 48 0 1 0 1 1]\n",
|
|
" [ 1 2 0 2 0 34 2 0 2 1]\n",
|
|
" [ 0 0 0 0 0 1 50 0 0 0]\n",
|
|
" [ 0 2 1 0 0 0 0 37 1 2]\n",
|
|
" [ 0 2 1 3 1 2 0 1 45 0]\n",
|
|
" [ 0 1 1 0 2 0 0 3 0 35]]\n",
|
|
"Zero-one classification loss :\n",
|
|
" 0.09399999999999997\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"####### Premier modèle de Classifier #######\n",
|
|
"\n",
|
|
"#Entraîne le classifier\n",
|
|
"clf = SVC(kernel=\"linear\")\n",
|
|
"# print(\"Training...\")\n",
|
|
"clf.fit(xtrain, ytrain)\n",
|
|
"\n",
|
|
"#Prédiction sur le jeu de tests\n",
|
|
"# print(\"Predicting...\")\n",
|
|
"t1 = time.time()\n",
|
|
"pred = clf.predict(xtest)\n",
|
|
"t2 = time.time()\n",
|
|
"print(\"Temps d'entraînement : \", round(t2-t1,5))\n",
|
|
"#print(\"Prédiction : \", pred)\n",
|
|
"# On calcule le score obtenu sur xtest avec les étiquettes ytest\n",
|
|
"score = clf.score(xtest, ytest)\n",
|
|
"print(\"Score échantillon de test : \", score)\n",
|
|
"\n",
|
|
"#Infos image 4\n",
|
|
"print(\"Classe image 4 : \", ytest[3])\n",
|
|
"print(\"Classe prédite image 4 : \", pred[3])\n",
|
|
"\n",
|
|
"#Calcul de différentes metrics\n",
|
|
"print(\"Précision pour chaque classe : \\n\", metrics.precision_score(ytest, pred,average=None))\n",
|
|
"print(\"Matrice de confusion SVM:\\n\", metrics.confusion_matrix(ytest, pred))\n",
|
|
"print(\"Zero-one classification loss :\\n\", metrics.zero_one_loss(ytest, pred))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 47,
|
|
"id": "5a4a5485",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Computing for kernel= poly ...\n",
|
|
"Computing for kernel= rbf ...\n",
|
|
"Computing for kernel= sigmoid ...\n",
|
|
"Computing for kernel= precomputed ...\n",
|
|
"Done\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"####### Variations de la fonction noyau #######\n",
|
|
"\n",
|
|
"list_training_times_kernel = []\n",
|
|
"list_precision_scores_kernel = []\n",
|
|
"list_zero_one_loss_kernel = []\n",
|
|
"\n",
|
|
"kernel_functions = [\"poly\",\"rbf\",\"sigmoid\",\"precomputed\"]\n",
|
|
"kernel_train = xtrain\n",
|
|
"kernel_test = xtest\n",
|
|
"for i in kernel_functions:\n",
|
|
" print(\"Computing for kernel=\", i, \"...\")\n",
|
|
" if (i == \"precomputed\"):\n",
|
|
" kernel_train=np.dot(xtrain,xtrain.T) # modified the train_set\n",
|
|
" kernel_test=np.dot(xtest,xtrain.T) # modified the test_set\n",
|
|
" \n",
|
|
" #Entraîne le classifier\n",
|
|
" clf = SVC(kernel=i)\n",
|
|
" t1 = round(time.time(),5)\n",
|
|
" clf.fit(kernel_train, ytrain)\n",
|
|
" t2 = round(time.time(),5)\n",
|
|
" #Prédiction sur le jeu de tests\n",
|
|
" pred = clf.predict(kernel_test)\n",
|
|
" # On sauvegarde le temps de calcul, la précision et \n",
|
|
" # les taux d'erreurs par classe\n",
|
|
" list_training_times_kernel.append(t2-t1)\n",
|
|
" list_precision_scores_kernel.append(clf.score(kernel_test, ytest))\n",
|
|
" list_zero_one_loss_kernel.append(metrics.zero_one_loss(ytest, pred))\n",
|
|
"print(\"Done\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 48,
|
|
"id": "9b961ed8",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"Text(36.0, 0.5, 'Zero-one loss')"
|
|
]
|
|
},
|
|
"execution_count": 48,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
},
|
|
{
|
|
"data": {
|
|
"image/png": "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\n",
|
|
"text/plain": [
|
|
"<Figure size 504x720 with 3 Axes>"
|
|
]
|
|
},
|
|
"metadata": {
|
|
"needs_background": "light"
|
|
},
|
|
"output_type": "display_data"
|
|
}
|
|
],
|
|
"source": [
|
|
"##### x coordinates\n",
|
|
"x = kernel_functions\n",
|
|
"training_times = list_training_times_kernel\n",
|
|
"precision_scores = list_precision_scores_kernel\n",
|
|
"zero_one_loss = list_zero_one_loss_kernel\n",
|
|
" \n",
|
|
"training_times = [round(i,2) for i in training_times]\n",
|
|
"precision_scores = [round(i,2) for i in precision_scores]\n",
|
|
"zero_one_loss = [round(i,2) for i in zero_one_loss]\n",
|
|
"\n",
|
|
"### Create plot\n",
|
|
"fig, figs = plt.subplots(nrows=3, ncols=1, figsize=(7,10))\n",
|
|
"fig.tight_layout(pad=3.0)\n",
|
|
"figs[0].plot(x,training_times, marker='o', color='r')\n",
|
|
"figs[1].plot(x,precision_scores, marker='o', color='b')\n",
|
|
"figs[2].plot(x,zero_one_loss, marker='o', color='g')\n",
|
|
"\n",
|
|
"### Add every x coordinates\n",
|
|
"figs[0].tick_params(axis='both', which='both', labelsize=7, labelbottom=True)\n",
|
|
"figs[1].tick_params(axis='both', which='both', labelsize=7, labelbottom=True)\n",
|
|
"figs[2].tick_params(axis='both', which='both', labelsize=7, labelbottom=True)\n",
|
|
"\n",
|
|
"for i in range(len(x)):\n",
|
|
" figs[0].annotate(training_times[i], # this is the text\n",
|
|
" (x[i],training_times[i]), # these are the coordinates to position the label\n",
|
|
" textcoords=\"offset points\", # how to position the text\n",
|
|
" xytext=(12,3), # distance from text to points (x,y)\n",
|
|
" ha='center') # horizontal alignment can be left, right or center\n",
|
|
" figs[1].annotate(precision_scores[i], # this is the text\n",
|
|
" (x[i],precision_scores[i]), # these are the coordinates to position the label\n",
|
|
" textcoords=\"offset points\", # how to position the text\n",
|
|
" xytext=(12,3), # distance from text to points (x,y)\n",
|
|
" ha='center') # horizontal alignment can be left, right or center\n",
|
|
" figs[2].annotate(zero_one_loss[i], # this is the text\n",
|
|
" (x[i],zero_one_loss[i]), # these are the coordinates to position the label\n",
|
|
" textcoords=\"offset points\", # how to position the text\n",
|
|
" xytext=(12,3), # distance from text to points (x,y)\n",
|
|
" ha='center') # horizontal alignment can be left, right or center\n",
|
|
"\n",
|
|
"figs[0].set_xticks(x)\n",
|
|
"figs[1].set_xticks(x)\n",
|
|
"figs[2].set_xticks(x)\n",
|
|
" \n",
|
|
"### Add title and axis names\n",
|
|
"figs[0].title.set_text('Training times for each kernel function')\n",
|
|
"figs[1].title.set_text('Precision score for each kernel function')\n",
|
|
"figs[2].title.set_text('Zero-one loss metrics for each kernel function')\n",
|
|
"figs[0].set_xlabel('kernel')\n",
|
|
"figs[1].set_xlabel('kernel')\n",
|
|
"figs[2].set_xlabel('kernel')\n",
|
|
"figs[0].set_ylabel('Training times (in seconds)')\n",
|
|
"figs[1].set_ylabel('Precision score')\n",
|
|
"figs[2].set_ylabel('Zero-one loss')"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 49,
|
|
"id": "5726fcb1",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Computing for C= 0.1 ...\n",
|
|
"Computing for C= 0.25 ...\n",
|
|
"Computing for C= 0.5 ...\n",
|
|
"Computing for C= 0.75 ...\n",
|
|
"Computing for C= 1.0 ...\n",
|
|
"Done\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"####### Variation du paramètre de tolérance aux erreurs C #######\n",
|
|
"\n",
|
|
"list_training_times_tol = []\n",
|
|
"list_precision_scores_tol = []\n",
|
|
"list_zero_one_loss_tol = []\n",
|
|
"\n",
|
|
"kernel_train = xtrain\n",
|
|
"kernel_test = xtest\n",
|
|
"tols = [0.1,0.25,0.5,0.75,1.0]\n",
|
|
"\n",
|
|
"for i in tols:\n",
|
|
" print(\"Computing for C=\", i, \"...\")\n",
|
|
" #Entraîne le classifier\n",
|
|
" clf = SVC(C=i, kernel=\"rbf\")\n",
|
|
" t1 = round(time.time(),5)\n",
|
|
" clf.fit(kernel_train, ytrain)\n",
|
|
" t2 = round(time.time(),5)\n",
|
|
" #Prédiction sur le jeu de tests\n",
|
|
" pred = clf.predict(kernel_test)\n",
|
|
" # On sauvegarde le temps de calcul, la précision et \n",
|
|
" # les taux d'erreurs par classe\n",
|
|
" list_training_times_tol.append(t2-t1)\n",
|
|
" list_precision_scores_tol.append(clf.score(kernel_test, ytest))\n",
|
|
" list_zero_one_loss_tol.append(metrics.zero_one_loss(ytest, pred))\n",
|
|
"print(\"Done\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 50,
|
|
"id": "741f82ca",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"Text(36.0, 0.5, 'Zero-one loss')"
|
|
]
|
|
},
|
|
"execution_count": 50,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
},
|
|
{
|
|
"data": {
|
|
"image/png": "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\n",
|
|
"text/plain": [
|
|
"<Figure size 504x720 with 3 Axes>"
|
|
]
|
|
},
|
|
"metadata": {
|
|
"needs_background": "light"
|
|
},
|
|
"output_type": "display_data"
|
|
}
|
|
],
|
|
"source": [
|
|
"##### x coordinates\n",
|
|
"x = tols\n",
|
|
"training_times = list_training_times_tol\n",
|
|
"precision_scores = list_precision_scores_tol\n",
|
|
"zero_one_loss = list_zero_one_loss_tol\n",
|
|
" \n",
|
|
"training_times = [round(i,2) for i in training_times]\n",
|
|
"precision_scores = [round(i,3) for i in precision_scores]\n",
|
|
"zero_one_loss = [round(i,3) for i in zero_one_loss]\n",
|
|
"\n",
|
|
"### Create plot\n",
|
|
"fig, figs = plt.subplots(nrows=3, ncols=1, figsize=(7,10))\n",
|
|
"fig.tight_layout(pad=3.0)\n",
|
|
"figs[0].plot(x,training_times, marker='o', color='r')\n",
|
|
"figs[1].plot(x,precision_scores, marker='o', color='b')\n",
|
|
"figs[2].plot(x,zero_one_loss, marker='o', color='g')\n",
|
|
"\n",
|
|
"### Add every x coordinates\n",
|
|
"figs[0].tick_params(axis='both', which='both', labelsize=7, labelbottom=True)\n",
|
|
"figs[1].tick_params(axis='both', which='both', labelsize=7, labelbottom=True)\n",
|
|
"figs[2].tick_params(axis='both', which='both', labelsize=7, labelbottom=True)\n",
|
|
"\n",
|
|
"for i in range(len(x)):\n",
|
|
" figs[0].annotate(training_times[i], # this is the text\n",
|
|
" (x[i],training_times[i]), # these are the coordinates to position the label\n",
|
|
" textcoords=\"offset points\", # how to position the text\n",
|
|
" xytext=(12,3), # distance from text to points (x,y)\n",
|
|
" ha='center') # horizontal alignment can be left, right or center\n",
|
|
" figs[1].annotate(precision_scores[i], # this is the text\n",
|
|
" (x[i],precision_scores[i]), # these are the coordinates to position the label\n",
|
|
" textcoords=\"offset points\", # how to position the text\n",
|
|
" xytext=(12,3), # distance from text to points (x,y)\n",
|
|
" ha='center') # horizontal alignment can be left, right or center\n",
|
|
" figs[2].annotate(zero_one_loss[i], # this is the text\n",
|
|
" (x[i],zero_one_loss[i]), # these are the coordinates to position the label\n",
|
|
" textcoords=\"offset points\", # how to position the text\n",
|
|
" xytext=(12,3), # distance from text to points (x,y)\n",
|
|
" ha='center') # horizontal alignment can be left, right or center\n",
|
|
"\n",
|
|
"figs[0].set_xticks(x)\n",
|
|
"figs[1].set_xticks(x)\n",
|
|
"figs[2].set_xticks(x)\n",
|
|
" \n",
|
|
"### Add title and axis names\n",
|
|
"figs[0].title.set_text('Training times for various level of tolerance (kernel=rbf)')\n",
|
|
"figs[1].title.set_text('Precision score for various level of tolerance (kernel=rbf)')\n",
|
|
"figs[2].title.set_text('Zero-one loss metrics various level of tolerance (kernel=rbf)')\n",
|
|
"figs[0].set_xlabel('tolerance')\n",
|
|
"figs[1].set_xlabel('tolerance')\n",
|
|
"figs[2].set_xlabel('tolerance')\n",
|
|
"figs[0].set_ylabel('Training times (in seconds)')\n",
|
|
"figs[1].set_ylabel('Precision score')\n",
|
|
"figs[2].set_ylabel('Zero-one loss')"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 51,
|
|
"id": "62c7302a",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Training...\n",
|
|
"Predicting...\n",
|
|
"Score échantillon de test : 0.9506666666666667\n",
|
|
"Précision pour chaque classe : \n",
|
|
" [0.99324324 0.97633136 0.93377483 0.95375723 0.92546584 0.91549296\n",
|
|
" 0.96527778 0.95833333 0.96402878 0.91472868]\n",
|
|
"Matrice de confusion pour C=1.0 et kernel=rbf :\n",
|
|
" [[147 0 0 0 1 0 0 0 1 0]\n",
|
|
" [ 0 165 0 0 0 0 0 0 1 0]\n",
|
|
" [ 1 2 141 0 0 1 1 1 0 0]\n",
|
|
" [ 0 0 3 165 0 3 0 1 1 1]\n",
|
|
" [ 0 0 1 0 149 0 1 0 0 5]\n",
|
|
" [ 0 0 0 4 2 130 3 1 1 1]\n",
|
|
" [ 0 0 1 0 0 3 139 0 0 0]\n",
|
|
" [ 0 1 3 0 4 0 0 138 0 4]\n",
|
|
" [ 0 1 1 3 1 4 0 0 134 0]\n",
|
|
" [ 0 0 1 1 4 1 0 3 1 118]]\n",
|
|
"Zero-one classification loss :\n",
|
|
" 0.04933333333333334\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"####### Meilleur modèle de SVM #######\n",
|
|
"\n",
|
|
"#Entraîne le classifier\n",
|
|
"clf = SVC(C=1.0,kernel=\"rbf\")\n",
|
|
"print(\"Training...\")\n",
|
|
"clf.fit(xtrain, ytrain)\n",
|
|
"\n",
|
|
"#Prédiction sur le jeu de tests\n",
|
|
"print(\"Predicting...\")\n",
|
|
"pred = clf.predict(xtest)\n",
|
|
"# On calcule le score obtenu sur xtest avec les étiquettes ytest\n",
|
|
"score = clf.score(xtest, ytest)\n",
|
|
"print(\"Score échantillon de test : \", score)\n",
|
|
"\n",
|
|
"#Calcul de différentes metrics\n",
|
|
"print(\"Précision pour chaque classe : \\n\", metrics.precision_score(ytest, pred,average=None))\n",
|
|
"print(\"Matrice de confusion pour C=1.0 et kernel=rbf :\\n\", metrics.confusion_matrix(ytest, pred))\n",
|
|
"print(\"Zero-one classification loss :\\n\", metrics.zero_one_loss(ytest, pred))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "30a232d5",
|
|
"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": "3.8.8"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|