diff --git a/naive_bayes.ipynb b/naive_bayes.ipynb index 6f5b26c..60c1a44 100644 --- a/naive_bayes.ipynb +++ b/naive_bayes.ipynb @@ -532,95 +532,17 @@ "The data of the features is continuous and non-binary. As such, we use a GaussianNB, the performance will nevertheless suffer as the features aren't all Normal distributed and the dimension is rather small, we cannot suppose normal distribution through size." ] }, - { - "cell_type": "markdown", - "id": "71cc45b2", - "metadata": {}, - "source": [ - "## Cross-Validation" - ] - }, - { - "cell_type": "code", - "execution_count": 7, - "id": "cff7e03a", - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Cross-validation (10 folds)\n", - "Average F1-score : 0.620 ± 0.217\n", - "Average accuracy : 0.655 ± 0.167\n", - "Average recall : 0.571 ± 0.262\n" - ] - }, - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], - "source": [ - "from sklearn.model_selection import cross_val_score\n", - "from sklearn.naive_bayes import GaussianNB\n", - "\n", - "general_classifier = GaussianNB()\n", - "general_classifier.fit(X_scaled, y)\n", - "# 10 fold corss-validation on the entire (transformed dataset)\n", - "f1_scores = cross_val_score(general_classifier, X_scaled, y, cv=10, scoring='f1')\n", - "acc_scores = cross_val_score(general_classifier, X_scaled, y, cv=10, scoring='accuracy')\n", - "recall_scores = cross_val_score(general_classifier, X_scaled, y, cv=10, scoring=\"recall\")\n", - "\n", - "# Report\n", - "print(\"Cross-validation (10 folds)\")\n", - "print(f\"Average F1-score : {f1_scores.mean():.3f} ± {f1_scores.std():.3f}\")\n", - "print(f\"Average accuracy : {acc_scores.mean():.3f} ± {acc_scores.std():.3f}\")\n", - "print(f\"Average recall : {recall_scores.mean():.3f} ± {recall_scores.std():.3f}\")\n", - "\n", - "# Visualisation des scores par fold\n", - "folds = range(1, len(f1_scores) + 1)\n", - "plt.figure(figsize=(8, 5))\n", - "plt.plot(folds, f1_scores, marker='o', label='F1-score')\n", - "plt.plot(folds, acc_scores, marker='s', label='Accuracy')\n", - "plt.plot(folds, recall_scores, marker=\"v\", label=\"Recall\")\n", - "plt.title(\"Scores by fold (cross-validation)\")\n", - "plt.xlabel(\"Fold\")\n", - "plt.ylabel(\"Score\")\n", - "plt.ylim(0, 1)\n", - "plt.grid(True)\n", - "plt.legend()\n", - "plt.tight_layout()\n", - "plt.show()\n" - ] - }, - { - "cell_type": "markdown", - "id": "94d0dd72", - "metadata": {}, - "source": [ - "### Analysis\n", - "\n", - "The high standard deviations shows a certain sensibilty to the folds, that is probably due to the small sample size. This high disparity between the folds also shows on the graph." - ] - }, { "cell_type": "markdown", "id": "5fd9b3f0", "metadata": {}, "source": [ - "## Final evaluation on the dataset" + "## Train/Test split for the evaluation " ] }, { "cell_type": "code", - "execution_count": 8, + "execution_count": null, "id": "d58417ad", "metadata": {}, "outputs": [ @@ -639,7 +561,7 @@ "# Stratified split to conserve the distribution of the variables\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X_scaled, y,\n", - " test_size=0.3,\n", + " test_size=0.2,\n", " random_state=42,\n", " stratify=y)\n", "\n", @@ -657,12 +579,77 @@ "- **Training set** : 81 observations\n", "- **Test set** : 24 observations\n", "\n", - "This is a standard 70/30 split" + "This is a standard 80/20 split" + ] + }, + { + "cell_type": "markdown", + "id": "13a5621f", + "metadata": {}, + "source": [ + "## Cross-Validation" ] }, { "cell_type": "code", - "execution_count": 9, + "execution_count": null, + "id": "3d50659c", + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn.model_selection import cross_val_score\n", + "from sklearn.naive_bayes import GaussianNB\n", + "\n", + "general_classifier = GaussianNB()\n", + "general_classifier.fit(X_train, y_train)\n", + "# 10 fold corss-validation on the entire (transformed dataset)\n", + "f1_scores = cross_val_score(general_classifier, X_train, y_train, cv=5, scoring='f1')\n", + "acc_scores = cross_val_score(general_classifier, X_train, y_train, cv=5, scoring='accuracy')\n", + "recall_scores = cross_val_score(general_classifier, X_train, y_train, cv=5, scoring=\"recall\")\n", + "\n", + "# Report\n", + "print(\"Cross-validation (5 folds)\")\n", + "print(f\"Average F1-score : {f1_scores.mean():.3f} ± {f1_scores.std():.3f}\")\n", + "print(f\"Average accuracy : {acc_scores.mean():.3f} ± {acc_scores.std():.3f}\")\n", + "print(f\"Average recall : {recall_scores.mean():.3f} ± {recall_scores.std():.3f}\")\n", + "\n", + "# Visualisation des scores par fold\n", + "folds = range(1, len(f1_scores) + 1)\n", + "plt.figure(figsize=(8, 5))\n", + "plt.plot(folds, f1_scores, marker='o', label='F1-score')\n", + "plt.plot(folds, acc_scores, marker='s', label='Accuracy')\n", + "plt.plot(folds, recall_scores, marker=\"v\", label=\"Recall\")\n", + "plt.title(\"Scores by fold (cross-validation)\")\n", + "plt.xlabel(\"Fold\")\n", + "plt.ylabel(\"Score\")\n", + "plt.ylim(0, 1)\n", + "plt.grid(True)\n", + "plt.legend()\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "c35db965", + "metadata": {}, + "source": [ + "### Analysis\n", + "\n", + "The high standard deviations shows a certain sensibilty to the folds, that is probably due to the small sample size. This high disparity between the folds also shows on the graph." + ] + }, + { + "cell_type": "markdown", + "id": "b8891271", + "metadata": {}, + "source": [ + "## Final evaluation on the dataset" + ] + }, + { + "cell_type": "code", + "execution_count": null, "id": "d26b8326", "metadata": {}, "outputs": [ @@ -708,7 +695,6 @@ "# Make predictions on the test set\n", "y_pred = nb_classifier.predict(X_test)\n", "\n", - "\n", "# Accuracy\n", "accuracy = accuracy_score(y_test, y_pred)\n", "print(f\"Accuracy : {accuracy:.3f}\")\n", diff --git a/neural_network.ipynb b/neural_network.ipynb index e4cd942..b57bdd1 100644 --- a/neural_network.ipynb +++ b/neural_network.ipynb @@ -191,9 +191,45 @@ " return model" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "train test split and scaling of the features " + ] + }, { "cell_type": "code", - "execution_count": 18, + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn.model_selection import train_test_split\n", + "from sklearn.preprocessing import StandardScaler\n", + "from sklearn.metrics import f1_score, classification_report\n", + "import tensorflow as tf\n", + "import numpy as np\n", + "\n", + "# Splitting the dataset into training and testing sets\n", + "\n", + "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n", + "\n", + "# Scaling the features\n", + "scaler = StandardScaler()\n", + "X_train_scaled = scaler.fit_transform(X_train)\n", + "X_test_scaled = scaler.transform(X_test)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Cross validation" + ] + }, + { + "cell_type": "code", + "execution_count": null, "metadata": {}, "outputs": [ { @@ -297,14 +333,9 @@ " verbose=1\n", ")\n", "\n", - "for fold, (train_idx, val_idx) in enumerate(skf.split(X, y), 1):\n", - " X_train, X_val = X.iloc[train_idx], X.iloc[val_idx]\n", - " y_train, y_val = y.iloc[train_idx], y.iloc[val_idx]\n", - "\n", - " # Standardisation\n", - " scaler = StandardScaler()\n", - " X_train_scaled = scaler.fit_transform(X_train)\n", - " X_val_scaled = scaler.transform(X_val)\n", + "for fold, (train_idx, val_idx) in enumerate(skf.split(X_train_scaled, y_train), 1):\n", + " X_cv_train, X_cv_val = X_train.iloc[train_idx], X_train.iloc[val_idx]\n", + " y_cv_train, y_cv_val = y_train.iloc[train_idx], y_train.iloc[val_idx]\n", " \n", " model = build_model()\n", "\n", @@ -319,10 +350,10 @@ "\n", " # Entraînement\n", " history = model.fit(\n", - " X_train_scaled, y_train,\n", + " X_cv_train, y_cv_train,\n", " epochs=50,\n", " batch_size=8,\n", - " validation_data=(X_val_scaled, y_val),\n", + " validation_data=(X_cv_val, y_cv_val),\n", " callbacks=[callback],\n", " verbose=0,\n", " class_weight={0: 1.0, 1: 2.0}\n", @@ -331,8 +362,8 @@ " histories.append(history.history)\n", "\n", " # Prédiction & F1\n", - " y_pred_val = (model.predict(X_val_scaled) > 0.5).astype(int)\n", - " score = f1_score(y_val, y_pred_val)\n", + " y_pred_val = (model.predict(X_cv_val) > 0.5).astype(int)\n", + " score = f1_score(y_cv_val, y_pred_val)\n", " f1_scores.append(score)\n", " print(f\"Fold {fold} - F1-score : {score:.4f}\")\n", "\n", @@ -400,7 +431,7 @@ }, { "cell_type": "code", - "execution_count": 20, + "execution_count": null, "metadata": {}, "outputs": [ { @@ -430,18 +461,6 @@ } ], "source": [ - "from sklearn.model_selection import train_test_split\n", - "from sklearn.preprocessing import StandardScaler\n", - "from sklearn.metrics import f1_score, classification_report\n", - "import tensorflow as tf\n", - "import numpy as np\n", - "\n", - "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n", - "\n", - "scaler = StandardScaler()\n", - "X_train_scaled = scaler.fit_transform(X_train)\n", - "X_test_scaled = scaler.transform(X_test)\n", - "\n", "model = build_model()\n", "\n", "model.compile(\n",