{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "provenance": [] }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "language_info": { "name": "python" } }, "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 1000 }, "id": "KTT5UnVuNqb7", "outputId": "f6469644-4a6d-48ac-b6d3-03dd0be9574f" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Clases: ['Alto' 'Bajo'] → encodings 0,1\n", "\n", "==================================================\n", " Regresión Logística\n", "==================================================\n", " Accuracy : 0.8433\n", " Precision: 0.8175 (Alto riesgo)\n", " Recall : 0.8110 (Alto riesgo)\n", " F1-Score : 0.8142 (Alto riesgo)\n", " AUC-ROC : 0.9293\n", " CV F1 : 0.8452\n", "\n", " Reporte completo:\n", " precision recall f1-score support\n", "\n", " Alto 0.82 0.81 0.81 127\n", " Bajo 0.86 0.87 0.86 173\n", "\n", " accuracy 0.84 300\n", " macro avg 0.84 0.84 0.84 300\n", "weighted avg 0.84 0.84 0.84 300\n", "\n", " Matriz de Confusión:\n", "[[103 24]\n", " [ 23 150]]\n", "\n", " Importancia de variables:\n", " Carga_Laboral: 1.2428\n", " Conflicto_Rol: 1.2159\n", " Estres: 1.2096\n", " Apoyo_Liderazgo: 1.1840\n", " Horas_Trabajo: 1.1804\n", "\n", "==================================================\n", " Random Forest\n", "==================================================\n", " Accuracy : 0.8200\n", " Precision: 0.7967 (Alto riesgo)\n", " Recall : 0.7717 (Alto riesgo)\n", " F1-Score : 0.7840 (Alto riesgo)\n", " AUC-ROC : 0.9099\n", " CV F1 : 0.8322\n", "\n", " Reporte completo:\n", " precision recall f1-score support\n", "\n", " Alto 0.80 0.77 0.78 127\n", " Bajo 0.84 0.86 0.85 173\n", "\n", " accuracy 0.82 300\n", " macro avg 0.82 0.81 0.81 300\n", "weighted avg 0.82 0.82 0.82 300\n", "\n", " Matriz de Confusión:\n", "[[ 98 29]\n", " [ 25 148]]\n", "\n", " Importancia de variables:\n", " Horas_Trabajo: 0.2655\n", " Carga_Laboral: 0.1916\n", " Apoyo_Liderazgo: 0.1835\n", " Estres: 0.1809\n", " Conflicto_Rol: 0.1785\n", "\n", "==================================================\n", " XGBoost\n", "==================================================\n", " Accuracy : 0.8233\n", " Precision: 0.7846 (Alto riesgo)\n", " Recall : 0.8031 (Alto riesgo)\n", " F1-Score : 0.7938 (Alto riesgo)\n", " AUC-ROC : 0.9093\n", " CV F1 : 0.8282\n", "\n", " Reporte completo:\n", " precision recall f1-score support\n", "\n", " Alto 0.78 0.80 0.79 127\n", " Bajo 0.85 0.84 0.85 173\n", "\n", " accuracy 0.82 300\n", " macro avg 0.82 0.82 0.82 300\n", "weighted avg 0.82 0.82 0.82 300\n", "\n", " Matriz de Confusión:\n", "[[102 25]\n", " [ 28 145]]\n", "\n", " Importancia de variables:\n", " Carga_Laboral: 0.2245\n", " Conflicto_Rol: 0.2098\n", " Estres: 0.2037\n", " Apoyo_Liderazgo: 0.2023\n", " Horas_Trabajo: 0.1596\n", "\n", "\n", "======================================================================\n", " TABLA COMPARATIVA FINAL\n", "======================================================================\n", "Modelo Accuracy Precision Recall F1 AUC\n", "----------------------------------------------------------------------\n", "Regresión Logística 0.8433 0.8175 0.8110 0.8142 0.9293\n", "Random Forest 0.8200 0.7967 0.7717 0.7840 0.9099\n", "XGBoost 0.8233 0.7846 0.8031 0.7938 0.9093\n", "Red Neuronal (MLP) 0.8300 0.8100 0.8400 0.8200 N/A \n" ] }, { "output_type": "error", "ename": "FileNotFoundError", "evalue": "[Errno 2] No such file or directory: '/home/claude/model_results.json'", "traceback": [ "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", "\u001b[0;31mFileNotFoundError\u001b[0m Traceback (most recent call last)", "\u001b[0;32m/tmp/ipykernel_18127/2722970189.py\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 132\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 133\u001b[0m \u001b[0;31m# Guardar JSON para usar en PPTX\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 134\u001b[0;31m \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'/home/claude/model_results.json'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'w'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 135\u001b[0m json.dump({\"results\": results, \"feature_importances\": feature_importances,\n\u001b[1;32m 136\u001b[0m \"classes\": le.classes_.tolist()}, f, indent=2)\n", "\u001b[0;31mFileNotFoundError\u001b[0m: [Errno 2] No such file or directory: '/home/claude/model_results.json'" ] } ], "source": [ "import pandas as pd\n", "import numpy as np\n", "import json\n", "from sklearn.model_selection import train_test_split, cross_val_score, StratifiedKFold\n", "from sklearn.preprocessing import StandardScaler, LabelEncoder\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.metrics import (accuracy_score, precision_score, recall_score,\n", " f1_score, roc_auc_score, confusion_matrix,\n", " classification_report)\n", "from xgboost import XGBClassifier\n", "import warnings\n", "warnings.filterwarnings('ignore')\n", "\n", "# ── 1. DATOS ──────────────────────────────────────────────\n", "df = pd.read_excel('/content/dataset_burnout_FINAL.xlsx',\n", " sheet_name='Dataset_Completo')\n", "\n", "top_variables = ['Carga_Laboral', 'Apoyo_Liderazgo', 'Conflicto_Rol', 'Estres', 'Horas_Trabajo']\n", "X = df[top_variables]\n", "y = df['Riesgo_Burnout']\n", "\n", "le = LabelEncoder()\n", "y_enc = le.fit_transform(y) # Alto=0, Bajo=1 → verificar\n", "print(\"Clases:\", le.classes_, \"→ encodings 0,1\")\n", "\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X, y_enc, test_size=0.2, random_state=42, stratify=y_enc)\n", "\n", "scaler = StandardScaler()\n", "X_train_s = scaler.fit_transform(X_train)\n", "X_test_s = scaler.transform(X_test)\n", "\n", "cv = StratifiedKFold(n_splits=5, shuffle=True, random_state=42)\n", "\n", "# ── 2. MODELOS ─────────────────────────────────────────────\n", "models = {\n", " \"Regresión Logística\": LogisticRegression(\n", " class_weight='balanced', max_iter=1000, random_state=42),\n", " \"Random Forest\": RandomForestClassifier(\n", " n_estimators=200, class_weight='balanced',\n", " max_depth=8, random_state=42, n_jobs=-1),\n", " \"XGBoost\": XGBClassifier(\n", " n_estimators=200, max_depth=5, learning_rate=0.05,\n", " scale_pos_weight=(y_enc==0).sum()/(y_enc==1).sum(),\n", " use_label_encoder=False, eval_metric='logloss',\n", " random_state=42, verbosity=0),\n", "}\n", "\n", "results = {}\n", "feature_importances = {}\n", "\n", "for name, model in models.items():\n", " print(f\"\\n{'='*50}\")\n", " print(f\" {name}\")\n", " print('='*50)\n", "\n", " # Usar datos escalados para RL, sin escalar para RF/XGB (ambos funcionan)\n", " Xtr = X_train_s if name == \"Regresión Logística\" else X_train_s\n", " Xte = X_test_s\n", "\n", " model.fit(Xtr, y_train)\n", " y_pred = model.predict(Xte)\n", " y_prob = model.predict_proba(Xte)[:, 1]\n", "\n", " # Métricas en test\n", " acc = accuracy_score(y_test, y_pred)\n", " prec = precision_score(y_test, y_pred, pos_label=0) # pos_label=0 = \"Alto\"\n", " rec = recall_score(y_test, y_pred, pos_label=0)\n", " f1 = f1_score(y_test, y_pred, pos_label=0)\n", " auc = roc_auc_score(y_test, y_prob)\n", " cv_sc = cross_val_score(model, X_train_s, y_train, cv=cv,\n", " scoring='f1', n_jobs=-1).mean()\n", "\n", " cm = confusion_matrix(y_test, y_pred)\n", "\n", " print(f\" Accuracy : {acc:.4f}\")\n", " print(f\" Precision: {prec:.4f} (Alto riesgo)\")\n", " print(f\" Recall : {rec:.4f} (Alto riesgo)\")\n", " print(f\" F1-Score : {f1:.4f} (Alto riesgo)\")\n", " print(f\" AUC-ROC : {auc:.4f}\")\n", " print(f\" CV F1 : {cv_sc:.4f}\")\n", " print(f\"\\n Reporte completo:\")\n", " print(classification_report(y_test, y_pred,\n", " target_names=[c for c in le.classes_]))\n", " print(f\" Matriz de Confusión:\\n{cm}\")\n", "\n", " results[name] = {\n", " \"accuracy\": round(acc, 4),\n", " \"precision_alto\": round(prec, 4),\n", " \"recall_alto\": round(rec, 4),\n", " \"f1_alto\": round(f1, 4),\n", " \"auc_roc\": round(auc, 4),\n", " \"cv_f1\": round(cv_sc, 4),\n", " \"cm\": cm.tolist(),\n", " }\n", "\n", " # Feature importance\n", " if hasattr(model, 'feature_importances_'):\n", " fi = dict(zip(top_variables, model.feature_importances_))\n", " elif hasattr(model, 'coef_'):\n", " fi = dict(zip(top_variables, np.abs(model.coef_[0])))\n", " else:\n", " fi = {}\n", " feature_importances[name] = fi\n", " if fi:\n", " print(\"\\n Importancia de variables:\")\n", " for var, imp in sorted(fi.items(), key=lambda x: -x[1]):\n", " print(f\" {var}: {imp:.4f}\")\n", "\n", "# Red Neuronal (resultados del notebook)\n", "results[\"Red Neuronal (MLP)\"] = {\n", " \"accuracy\": 0.83,\n", " \"precision_alto\": 0.81,\n", " \"recall_alto\": 0.84,\n", " \"f1_alto\": 0.82,\n", " \"auc_roc\": None,\n", " \"cv_f1\": None,\n", " \"cm\": None,\n", "}\n", "\n", "# ── 3. TABLA RESUMEN ──────────────────────────────────────\n", "print(\"\\n\\n\" + \"=\"*70)\n", "print(\" TABLA COMPARATIVA FINAL\")\n", "print(\"=\"*70)\n", "header = f\"{'Modelo':<28} {'Accuracy':>9} {'Precision':>10} {'Recall':>8} {'F1':>8} {'AUC':>8}\"\n", "print(header)\n", "print(\"-\"*70)\n", "for name, r in results.items():\n", " auc_str = f\"{r['auc_roc']:.4f}\" if r['auc_roc'] else \" N/A \"\n", " print(f\"{name:<28} {r['accuracy']:>9.4f} {r['precision_alto']:>10.4f} {r['recall_alto']:>8.4f} {r['f1_alto']:>8.4f} {auc_str:>8}\")\n", "\n", "# Guardar JSON para usar en PPTX\n", "with open('/home/claude/model_results.json', 'w') as f:\n", " json.dump({\"results\": results, \"feature_importances\": feature_importances,\n", " \"classes\": le.classes_.tolist()}, f, indent=2)\n", "\n", "print(\"\\n✅ Resultados guardados en model_results.json\")" ] } ] }