{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clasificador Multiclase — Dry Bean Dataset\n",
    "\n",
    "**Curso:** Fundamentos de Machine Learning  \n",
    "**Nombre:** Alejandro Lembke Barrientos | Carné: 12002840  \n",
    "**Dataset:** Dry Bean Dataset (UCI ID 602) — 7 clases, 16 features, 13,611 instancias"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Setup & Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "b18469ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from ucimlrepo import fetch_ucirepo\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import precision_score, recall_score, f1_score, accuracy_score\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from xgboost import XGBClassifier\n",
    "from scipy.stats import mode\n",
    "\n",
    "sys.path.append(os.path.join(os.getcwd(), 'src'))\n",
    "from utils import metricas, registrar, plot_importancia\n",
    "\n",
    "RANDOM_STATE = 42\n",
    "np.random.seed(RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41902be2",
   "metadata": {},
   "source": [
    "## 2. Carga y Exploración del Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "f264ed2e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (13611, 17)\n",
      "\n",
      "Metadata:\n",
      "Images of 13,611 grains of 7 different registered dry beans were taken with a high-resolution camera. A total of 16 features; 12 dimensions and 4 shape forms, were obtained from the grains.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Area</th>\n",
       "      <th>Perimeter</th>\n",
       "      <th>MajorAxisLength</th>\n",
       "      <th>MinorAxisLength</th>\n",
       "      <th>AspectRatio</th>\n",
       "      <th>Eccentricity</th>\n",
       "      <th>ConvexArea</th>\n",
       "      <th>EquivDiameter</th>\n",
       "      <th>Extent</th>\n",
       "      <th>Solidity</th>\n",
       "      <th>Roundness</th>\n",
       "      <th>Compactness</th>\n",
       "      <th>ShapeFactor1</th>\n",
       "      <th>ShapeFactor2</th>\n",
       "      <th>ShapeFactor3</th>\n",
       "      <th>ShapeFactor4</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>28395</td>\n",
       "      <td>610.291</td>\n",
       "      <td>208.178117</td>\n",
       "      <td>173.888747</td>\n",
       "      <td>1.197191</td>\n",
       "      <td>0.549812</td>\n",
       "      <td>28715</td>\n",
       "      <td>190.141097</td>\n",
       "      <td>0.763923</td>\n",
       "      <td>0.988856</td>\n",
       "      <td>0.958027</td>\n",
       "      <td>0.913358</td>\n",
       "      <td>0.007332</td>\n",
       "      <td>0.003147</td>\n",
       "      <td>0.834222</td>\n",
       "      <td>0.998724</td>\n",
       "      <td>SEKER</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>28734</td>\n",
       "      <td>638.018</td>\n",
       "      <td>200.524796</td>\n",
       "      <td>182.734419</td>\n",
       "      <td>1.097356</td>\n",
       "      <td>0.411785</td>\n",
       "      <td>29172</td>\n",
       "      <td>191.272751</td>\n",
       "      <td>0.783968</td>\n",
       "      <td>0.984986</td>\n",
       "      <td>0.887034</td>\n",
       "      <td>0.953861</td>\n",
       "      <td>0.006979</td>\n",
       "      <td>0.003564</td>\n",
       "      <td>0.909851</td>\n",
       "      <td>0.998430</td>\n",
       "      <td>SEKER</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>29380</td>\n",
       "      <td>624.110</td>\n",
       "      <td>212.826130</td>\n",
       "      <td>175.931143</td>\n",
       "      <td>1.209713</td>\n",
       "      <td>0.562727</td>\n",
       "      <td>29690</td>\n",
       "      <td>193.410904</td>\n",
       "      <td>0.778113</td>\n",
       "      <td>0.989559</td>\n",
       "      <td>0.947849</td>\n",
       "      <td>0.908774</td>\n",
       "      <td>0.007244</td>\n",
       "      <td>0.003048</td>\n",
       "      <td>0.825871</td>\n",
       "      <td>0.999066</td>\n",
       "      <td>SEKER</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>30008</td>\n",
       "      <td>645.884</td>\n",
       "      <td>210.557999</td>\n",
       "      <td>182.516516</td>\n",
       "      <td>1.153638</td>\n",
       "      <td>0.498616</td>\n",
       "      <td>30724</td>\n",
       "      <td>195.467062</td>\n",
       "      <td>0.782681</td>\n",
       "      <td>0.976696</td>\n",
       "      <td>0.903936</td>\n",
       "      <td>0.928329</td>\n",
       "      <td>0.007017</td>\n",
       "      <td>0.003215</td>\n",
       "      <td>0.861794</td>\n",
       "      <td>0.994199</td>\n",
       "      <td>SEKER</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>30140</td>\n",
       "      <td>620.134</td>\n",
       "      <td>201.847882</td>\n",
       "      <td>190.279279</td>\n",
       "      <td>1.060798</td>\n",
       "      <td>0.333680</td>\n",
       "      <td>30417</td>\n",
       "      <td>195.896503</td>\n",
       "      <td>0.773098</td>\n",
       "      <td>0.990893</td>\n",
       "      <td>0.984877</td>\n",
       "      <td>0.970516</td>\n",
       "      <td>0.006697</td>\n",
       "      <td>0.003665</td>\n",
       "      <td>0.941900</td>\n",
       "      <td>0.999166</td>\n",
       "      <td>SEKER</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Area  Perimeter  MajorAxisLength  MinorAxisLength  AspectRatio  \\\n",
       "0  28395    610.291       208.178117       173.888747     1.197191   \n",
       "1  28734    638.018       200.524796       182.734419     1.097356   \n",
       "2  29380    624.110       212.826130       175.931143     1.209713   \n",
       "3  30008    645.884       210.557999       182.516516     1.153638   \n",
       "4  30140    620.134       201.847882       190.279279     1.060798   \n",
       "\n",
       "   Eccentricity  ConvexArea  EquivDiameter    Extent  Solidity  Roundness  \\\n",
       "0      0.549812       28715     190.141097  0.763923  0.988856   0.958027   \n",
       "1      0.411785       29172     191.272751  0.783968  0.984986   0.887034   \n",
       "2      0.562727       29690     193.410904  0.778113  0.989559   0.947849   \n",
       "3      0.498616       30724     195.467062  0.782681  0.976696   0.903936   \n",
       "4      0.333680       30417     195.896503  0.773098  0.990893   0.984877   \n",
       "\n",
       "   Compactness  ShapeFactor1  ShapeFactor2  ShapeFactor3  ShapeFactor4  Class  \n",
       "0     0.913358      0.007332      0.003147      0.834222      0.998724  SEKER  \n",
       "1     0.953861      0.006979      0.003564      0.909851      0.998430  SEKER  \n",
       "2     0.908774      0.007244      0.003048      0.825871      0.999066  SEKER  \n",
       "3     0.928329      0.007017      0.003215      0.861794      0.994199  SEKER  \n",
       "4     0.970516      0.006697      0.003665      0.941900      0.999166  SEKER  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dry_bean = fetch_ucirepo(id=602)\n",
    "\n",
    "X_df   = dry_bean.data.features          # DataFrame 13611 x 16\n",
    "y_raw  = dry_bean.data.targets.squeeze() # Series con etiquetas de clase\n",
    "\n",
    "df = X_df.copy()\n",
    "df['Class'] = y_raw.values\n",
    "\n",
    "print('Shape:', df.shape)\n",
    "print('\\nMetadata:')\n",
    "print(dry_bean.metadata['abstract'])\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "13e1a95a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Valores nulos por columna:\n",
      "Area               0\n",
      "Perimeter          0\n",
      "MajorAxisLength    0\n",
      "MinorAxisLength    0\n",
      "AspectRatio        0\n",
      "Eccentricity       0\n",
      "ConvexArea         0\n",
      "EquivDiameter      0\n",
      "Extent             0\n",
      "Solidity           0\n",
      "Roundness          0\n",
      "Compactness        0\n",
      "ShapeFactor1       0\n",
      "ShapeFactor2       0\n",
      "ShapeFactor3       0\n",
      "ShapeFactor4       0\n",
      "Class              0\n",
      "dtype: int64\n",
      "\n",
      "Distribución de clases:\n",
      "Class\n",
      "DERMASON    3546\n",
      "SIRA        2636\n",
      "SEKER       2027\n",
      "HOROZ       1928\n",
      "CALI        1630\n",
      "BARBUNYA    1322\n",
      "BOMBAY       522\n",
      "Name: count, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print('Valores nulos por columna:')\n",
    "print(df.isnull().sum())\n",
    "print('\\nDistribución de clases:')\n",
    "print(df['Class'].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "caa5998f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Area</th>\n",
       "      <th>Perimeter</th>\n",
       "      <th>MajorAxisLength</th>\n",
       "      <th>MinorAxisLength</th>\n",
       "      <th>AspectRatio</th>\n",
       "      <th>Eccentricity</th>\n",
       "      <th>ConvexArea</th>\n",
       "      <th>EquivDiameter</th>\n",
       "      <th>Extent</th>\n",
       "      <th>Solidity</th>\n",
       "      <th>Roundness</th>\n",
       "      <th>Compactness</th>\n",
       "      <th>ShapeFactor1</th>\n",
       "      <th>ShapeFactor2</th>\n",
       "      <th>ShapeFactor3</th>\n",
       "      <th>ShapeFactor4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "      <td>13611.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>53048.284549</td>\n",
       "      <td>855.283459</td>\n",
       "      <td>320.141867</td>\n",
       "      <td>202.270714</td>\n",
       "      <td>1.583242</td>\n",
       "      <td>0.750895</td>\n",
       "      <td>53768.200206</td>\n",
       "      <td>253.064220</td>\n",
       "      <td>0.749733</td>\n",
       "      <td>0.987143</td>\n",
       "      <td>0.873282</td>\n",
       "      <td>0.799864</td>\n",
       "      <td>0.006564</td>\n",
       "      <td>0.001716</td>\n",
       "      <td>0.643590</td>\n",
       "      <td>0.995063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>29324.095717</td>\n",
       "      <td>214.289696</td>\n",
       "      <td>85.694186</td>\n",
       "      <td>44.970091</td>\n",
       "      <td>0.246678</td>\n",
       "      <td>0.092002</td>\n",
       "      <td>29774.915817</td>\n",
       "      <td>59.177120</td>\n",
       "      <td>0.049086</td>\n",
       "      <td>0.004660</td>\n",
       "      <td>0.059520</td>\n",
       "      <td>0.061713</td>\n",
       "      <td>0.001128</td>\n",
       "      <td>0.000596</td>\n",
       "      <td>0.098996</td>\n",
       "      <td>0.004366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>20420.000000</td>\n",
       "      <td>524.736000</td>\n",
       "      <td>183.601165</td>\n",
       "      <td>122.512653</td>\n",
       "      <td>1.024868</td>\n",
       "      <td>0.218951</td>\n",
       "      <td>20684.000000</td>\n",
       "      <td>161.243764</td>\n",
       "      <td>0.555315</td>\n",
       "      <td>0.919246</td>\n",
       "      <td>0.489618</td>\n",
       "      <td>0.640577</td>\n",
       "      <td>0.002778</td>\n",
       "      <td>0.000564</td>\n",
       "      <td>0.410339</td>\n",
       "      <td>0.947687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>36328.000000</td>\n",
       "      <td>703.523500</td>\n",
       "      <td>253.303633</td>\n",
       "      <td>175.848170</td>\n",
       "      <td>1.432307</td>\n",
       "      <td>0.715928</td>\n",
       "      <td>36714.500000</td>\n",
       "      <td>215.068003</td>\n",
       "      <td>0.718634</td>\n",
       "      <td>0.985670</td>\n",
       "      <td>0.832096</td>\n",
       "      <td>0.762469</td>\n",
       "      <td>0.005900</td>\n",
       "      <td>0.001154</td>\n",
       "      <td>0.581359</td>\n",
       "      <td>0.993703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>44652.000000</td>\n",
       "      <td>794.941000</td>\n",
       "      <td>296.883367</td>\n",
       "      <td>192.431733</td>\n",
       "      <td>1.551124</td>\n",
       "      <td>0.764441</td>\n",
       "      <td>45178.000000</td>\n",
       "      <td>238.438026</td>\n",
       "      <td>0.759859</td>\n",
       "      <td>0.988283</td>\n",
       "      <td>0.883157</td>\n",
       "      <td>0.801277</td>\n",
       "      <td>0.006645</td>\n",
       "      <td>0.001694</td>\n",
       "      <td>0.642044</td>\n",
       "      <td>0.996386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>61332.000000</td>\n",
       "      <td>977.213000</td>\n",
       "      <td>376.495012</td>\n",
       "      <td>217.031741</td>\n",
       "      <td>1.707109</td>\n",
       "      <td>0.810466</td>\n",
       "      <td>62294.000000</td>\n",
       "      <td>279.446467</td>\n",
       "      <td>0.786851</td>\n",
       "      <td>0.990013</td>\n",
       "      <td>0.916869</td>\n",
       "      <td>0.834270</td>\n",
       "      <td>0.007271</td>\n",
       "      <td>0.002170</td>\n",
       "      <td>0.696006</td>\n",
       "      <td>0.997883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>254616.000000</td>\n",
       "      <td>1985.370000</td>\n",
       "      <td>738.860154</td>\n",
       "      <td>460.198497</td>\n",
       "      <td>2.430306</td>\n",
       "      <td>0.911423</td>\n",
       "      <td>263261.000000</td>\n",
       "      <td>569.374358</td>\n",
       "      <td>0.866195</td>\n",
       "      <td>0.994677</td>\n",
       "      <td>0.990685</td>\n",
       "      <td>0.987303</td>\n",
       "      <td>0.010451</td>\n",
       "      <td>0.003665</td>\n",
       "      <td>0.974767</td>\n",
       "      <td>0.999733</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Area     Perimeter  MajorAxisLength  MinorAxisLength  \\\n",
       "count   13611.000000  13611.000000     13611.000000     13611.000000   \n",
       "mean    53048.284549    855.283459       320.141867       202.270714   \n",
       "std     29324.095717    214.289696        85.694186        44.970091   \n",
       "min     20420.000000    524.736000       183.601165       122.512653   \n",
       "25%     36328.000000    703.523500       253.303633       175.848170   \n",
       "50%     44652.000000    794.941000       296.883367       192.431733   \n",
       "75%     61332.000000    977.213000       376.495012       217.031741   \n",
       "max    254616.000000   1985.370000       738.860154       460.198497   \n",
       "\n",
       "        AspectRatio  Eccentricity     ConvexArea  EquivDiameter        Extent  \\\n",
       "count  13611.000000  13611.000000   13611.000000   13611.000000  13611.000000   \n",
       "mean       1.583242      0.750895   53768.200206     253.064220      0.749733   \n",
       "std        0.246678      0.092002   29774.915817      59.177120      0.049086   \n",
       "min        1.024868      0.218951   20684.000000     161.243764      0.555315   \n",
       "25%        1.432307      0.715928   36714.500000     215.068003      0.718634   \n",
       "50%        1.551124      0.764441   45178.000000     238.438026      0.759859   \n",
       "75%        1.707109      0.810466   62294.000000     279.446467      0.786851   \n",
       "max        2.430306      0.911423  263261.000000     569.374358      0.866195   \n",
       "\n",
       "           Solidity     Roundness   Compactness  ShapeFactor1  ShapeFactor2  \\\n",
       "count  13611.000000  13611.000000  13611.000000  13611.000000  13611.000000   \n",
       "mean       0.987143      0.873282      0.799864      0.006564      0.001716   \n",
       "std        0.004660      0.059520      0.061713      0.001128      0.000596   \n",
       "min        0.919246      0.489618      0.640577      0.002778      0.000564   \n",
       "25%        0.985670      0.832096      0.762469      0.005900      0.001154   \n",
       "50%        0.988283      0.883157      0.801277      0.006645      0.001694   \n",
       "75%        0.990013      0.916869      0.834270      0.007271      0.002170   \n",
       "max        0.994677      0.990685      0.987303      0.010451      0.003665   \n",
       "\n",
       "       ShapeFactor3  ShapeFactor4  \n",
       "count  13611.000000  13611.000000  \n",
       "mean       0.643590      0.995063  \n",
       "std        0.098996      0.004366  \n",
       "min        0.410339      0.947687  \n",
       "25%        0.581359      0.993703  \n",
       "50%        0.642044      0.996386  \n",
       "75%        0.696006      0.997883  \n",
       "max        0.974767      0.999733  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "92e35ff2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Distribución de clases\n",
    "df['Class'].value_counts().plot(kind='bar', title='Distribución de Clases')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "caca2128",
   "metadata": {},
   "source": [
    "## 3. Preparación de Datos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "b19d4476",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clases: ['BARBUNYA' 'BOMBAY' 'CALI' 'DERMASON' 'HOROZ' 'SEKER' 'SIRA']\n",
      "Train: (10888, 16) | Test: (2723, 16)\n",
      "Escalado aplicado — X_train_scaled shape: (10888, 16)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "feature_names = [c for c in df.columns if c != 'Class']\n",
    "X = df[feature_names].values\n",
    "y_raw = df['Class'].values\n",
    "\n",
    "le = LabelEncoder()\n",
    "y = le.fit_transform(y_raw)\n",
    "print('Clases:', le.classes_)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.2, stratify=y, random_state=RANDOM_STATE\n",
    ")\n",
    "print(f'Train: {X_train.shape} | Test: {X_test.shape}')\n",
    "\n",
    "# Escalado para la Red Neuronal (fit solo en train, transform en ambos)\n",
    "scaler = StandardScaler()\n",
    "X_train_scaled = scaler.fit_transform(X_train)\n",
    "X_test_scaled  = scaler.transform(X_test)\n",
    "print('Escalado aplicado — X_train_scaled shape:', X_train_scaled.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ce7295c",
   "metadata": {},
   "source": [
    "## 4. Bitácora — Funciones\n",
    "\n",
    "Las funciones `metricas()` y `registrar()` están en `src/utils.py`.  \n",
    "Registran cada experimento en `outputs/bitacora_experimentos.csv` en modo append."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f11cba6f",
   "metadata": {},
   "source": [
    "## 5. Experimentos — Random Forest\n",
    "\n",
    "≥5 experimentos variando: `n_estimators`, `max_depth`, `min_samples_leaf`, `max_features`, `criterion`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "7705ae9c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[bitacora] random_forest | RF_01 | F1 test=0.9328\n",
      "[bitacora] random_forest | RF_02 | F1 test=0.9319\n",
      "[bitacora] random_forest | RF_03 | F1 test=0.9311\n",
      "[bitacora] random_forest | RF_04 | F1 test=0.9317\n",
      "[bitacora] random_forest | RF_05 | F1 test=0.9316\n",
      "Experimentos RF registrados.\n"
     ]
    }
   ],
   "source": [
    "experimentos_rf = [\n",
    "    {'n_estimators': 100, 'max_depth': None, 'min_samples_leaf': 1,  'max_features': 'sqrt', 'criterion': 'gini'},\n",
    "    {'n_estimators': 200, 'max_depth': 10,   'min_samples_leaf': 1,  'max_features': 'sqrt', 'criterion': 'gini'},\n",
    "    {'n_estimators': 200, 'max_depth': 20,   'min_samples_leaf': 2,  'max_features': 'sqrt', 'criterion': 'gini'},\n",
    "    {'n_estimators': 500, 'max_depth': None, 'min_samples_leaf': 5,  'max_features': 'log2', 'criterion': 'gini'},\n",
    "    {'n_estimators': 300, 'max_depth': None, 'min_samples_leaf': 1,  'max_features': 'sqrt', 'criterion': 'entropy'},\n",
    "]\n",
    "\n",
    "for i, cfg in enumerate(experimentos_rf, 1):\n",
    "    model = RandomForestClassifier(**cfg, random_state=RANDOM_STATE, n_jobs=-1)\n",
    "    model.fit(X_train, y_train)\n",
    "    registrar(f'RF_{i:02d}', 'random_forest', cfg, '', model, X_train, y_train, X_test, y_test)\n",
    "\n",
    "print('Experimentos RF registrados.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "45162343",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  experimento_id  f1_train   f1_test\n",
      "0          RF_01  1.000000  0.932772\n",
      "1          RF_02  0.970755  0.931902\n",
      "3          RF_04  0.968360  0.931652\n",
      "4          RF_05  1.000000  0.931572\n",
      "2          RF_03  0.989488  0.931058\n",
      "\n",
      "Mejor RF: {'n_estimators': 100, 'max_depth': None, 'min_samples_leaf': 1, 'max_features': 'sqrt', 'criterion': 'gini'}\n"
     ]
    }
   ],
   "source": [
    "# Leer bitácora y seleccionar mejor RF\n",
    "df_bit = pd.read_csv(os.path.join('outputs', 'bitacora_experimentos.csv'))\n",
    "df_rf = df_bit[df_bit['modelo'] == 'random_forest']\n",
    "print(df_rf[['experimento_id', 'f1_train', 'f1_test']].sort_values('f1_test', ascending=False))\n",
    "\n",
    "mejor_rf_cfg = experimentos_rf[df_rf['f1_test'].idxmax() - df_rf.index[0]]\n",
    "mejor_rf = RandomForestClassifier(**mejor_rf_cfg, random_state=RANDOM_STATE, n_jobs=-1)\n",
    "mejor_rf.fit(X_train, y_train)\n",
    "print('\\nMejor RF:', mejor_rf_cfg)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14a08696",
   "metadata": {},
   "source": [
    "## 6. Experimentos — XGBoost\n",
    "\n",
    "≥5 experimentos variando: `n_estimators`, `max_depth`, `learning_rate`, `subsample`, `colsample_bytree`, `reg_lambda`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "3cdc3f32",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[bitacora] xgboost | XGB_01 | F1 test=0.9363\n",
      "[bitacora] xgboost | XGB_02 | F1 test=0.9377\n",
      "[bitacora] xgboost | XGB_03 | F1 test=0.9370\n",
      "[bitacora] xgboost | XGB_04 | F1 test=0.9340\n",
      "[bitacora] xgboost | XGB_05 | F1 test=0.9365\n",
      "Experimentos XGBoost registrados.\n"
     ]
    }
   ],
   "source": [
    "experimentos_xgb = [\n",
    "    {'n_estimators': 100, 'max_depth': 6,  'learning_rate': 0.3,  'subsample': 1.0, 'colsample_bytree': 1.0, 'reg_lambda': 1},\n",
    "    {'n_estimators': 200, 'max_depth': 6,  'learning_rate': 0.1,  'subsample': 0.8, 'colsample_bytree': 0.8, 'reg_lambda': 1},\n",
    "    {'n_estimators': 300, 'max_depth': 4,  'learning_rate': 0.05, 'subsample': 0.8, 'colsample_bytree': 0.8, 'reg_lambda': 2},\n",
    "    {'n_estimators': 200, 'max_depth': 8,  'learning_rate': 0.1,  'subsample': 0.6, 'colsample_bytree': 0.6, 'reg_lambda': 5},\n",
    "    {'n_estimators': 500, 'max_depth': 6,  'learning_rate': 0.01, 'subsample': 0.8, 'colsample_bytree': 0.8, 'reg_lambda': 1},\n",
    "]\n",
    "\n",
    "for i, cfg in enumerate(experimentos_xgb, 1):\n",
    "    model = XGBClassifier(**cfg, random_state=RANDOM_STATE, eval_metric='mlogloss', verbosity=0)\n",
    "    model.fit(X_train, y_train)\n",
    "    registrar(f'XGB_{i:02d}', 'xgboost', cfg, '', model, X_train, y_train, X_test, y_test)\n",
    "\n",
    "print('Experimentos XGBoost registrados.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "1472d6b1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  experimento_id  f1_train   f1_test\n",
      "6         XGB_02  0.999447  0.937734\n",
      "7         XGB_03  0.972587  0.936978\n",
      "9         XGB_05  0.969759  0.936526\n",
      "5         XGB_01  0.999941  0.936272\n",
      "8         XGB_04  0.997685  0.933975\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Mejor XGBoost: {'n_estimators': 200, 'max_depth': 6, 'learning_rate': 0.1, 'subsample': 0.8, 'colsample_bytree': 0.8, 'reg_lambda': 1}\n"
     ]
    }
   ],
   "source": [
    "df_bit = pd.read_csv(os.path.join('outputs', 'bitacora_experimentos.csv'))\n",
    "df_xgb = df_bit[df_bit['modelo'] == 'xgboost']\n",
    "print(df_xgb[['experimento_id', 'f1_train', 'f1_test']].sort_values('f1_test', ascending=False))\n",
    "\n",
    "mejor_xgb_cfg = experimentos_xgb[df_xgb['f1_test'].idxmax() - df_xgb.index[0]]\n",
    "mejor_xgb = XGBClassifier(**mejor_xgb_cfg, random_state=RANDOM_STATE, eval_metric='mlogloss', verbosity=0)\n",
    "mejor_xgb.fit(X_train, y_train)\n",
    "print('\\nMejor XGBoost:', mejor_xgb_cfg)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cbfd6199",
   "metadata": {},
   "source": [
    "## 7. Experimentos — Red Neuronal (Keras)\n",
    "\n",
    "≥5 experimentos variando: capas, neuronas, `dropout`, `lr`, `batch_size`, `epochs`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "6fd7b0c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
      "I0000 00:00:1782283235.163821  678660 cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.\n",
      "I0000 00:00:1782283235.560729  678660 cpu_feature_guard.cc:227] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
      "To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
      "I0000 00:00:1782283237.375377  678660 cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.\n",
      "E0000 00:00:1782283237.662609  678660 cuda_platform.cc:52] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)\n",
      "I0000 00:00:1782283237.662748  678660 cuda_diagnostics.cc:171] verbose logging is disabled. Rerun with verbose logging (usually --v=1 or --vmodule=cuda_diagnostics=1) to get more diagnostic output from this module\n",
      "I0000 00:00:1782283237.662754  678660 cuda_diagnostics.cc:176] retrieving CUDA diagnostic information for host: 41ad1eab7f7e\n",
      "I0000 00:00:1782283237.662760  678660 cuda_diagnostics.cc:183] hostname: 41ad1eab7f7e\n",
      "I0000 00:00:1782283237.662891  678660 cuda_diagnostics.cc:190] libcuda reported version is: NOT_FOUND: was unable to find libcuda.so DSO loaded into this program. The library may be missing or provided via another object.\n",
      "I0000 00:00:1782283237.663023  678660 cuda_diagnostics.cc:194] kernel reported version is: 580.159.3\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[bitacora] neural_network | NN_01 | F1 test=0.9374\n",
      "[bitacora] neural_network | NN_02 | F1 test=0.9408\n",
      "[bitacora] neural_network | NN_03 | F1 test=0.9388\n",
      "[bitacora] neural_network | NN_04 | F1 test=0.9386\n",
      "[bitacora] neural_network | NN_05 | F1 test=0.9383\n",
      "Experimentos NN (Keras) registrados.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "\n",
    "N_CLASES   = len(le.classes_)        # 7\n",
    "N_FEATURES = X_train_scaled.shape[1] # 16\n",
    "\n",
    "def crear_modelo(capas, activacion='relu', dropout=0.0, lr=0.001):\n",
    "    model = keras.Sequential()\n",
    "    model.add(keras.Input(shape=(N_FEATURES,)))\n",
    "    for neuronas in capas:\n",
    "        model.add(layers.Dense(neuronas, activation=activacion))\n",
    "        if dropout > 0:\n",
    "            model.add(layers.Dropout(dropout))\n",
    "    model.add(layers.Dense(N_CLASES, activation='softmax'))\n",
    "    model.compile(\n",
    "        optimizer=keras.optimizers.Adam(learning_rate=lr),\n",
    "        loss='sparse_categorical_crossentropy',\n",
    "    )\n",
    "    return model\n",
    "\n",
    "\n",
    "class KerasWrapper:\n",
    "    \"\"\"Wrapper sklearn-compatible para modelos Keras.\"\"\"\n",
    "    def __init__(self, model, batch_size=64, epochs=100):\n",
    "        self.model      = model\n",
    "        self.batch_size = batch_size\n",
    "        self.epochs     = epochs\n",
    "\n",
    "    def predict(self, X):\n",
    "        return np.argmax(self.model.predict(X, verbose=0), axis=1)\n",
    "\n",
    "    def fit(self, X, y, **kwargs):\n",
    "        self.model.fit(X, y,\n",
    "                       batch_size=self.batch_size,\n",
    "                       epochs=self.epochs,\n",
    "                       verbose=0)\n",
    "        return self\n",
    "\n",
    "\n",
    "experimentos_nn = [\n",
    "    {'capas': (64,),          'activacion': 'relu', 'dropout': 0.0, 'lr': 0.001,  'batch_size': 32,  'epochs': 50,  'arquitectura': '[64]→softmax'},\n",
    "    {'capas': (128, 64),      'activacion': 'relu', 'dropout': 0.3, 'lr': 0.001,  'batch_size': 64,  'epochs': 100, 'arquitectura': '[128,64]+Dropout(0.3)→softmax'},\n",
    "    {'capas': (256, 128, 64), 'activacion': 'relu', 'dropout': 0.3, 'lr': 0.001,  'batch_size': 64,  'epochs': 100, 'arquitectura': '[256,128,64]+Dropout(0.3)→softmax'},\n",
    "    {'capas': (128, 64),      'activacion': 'relu', 'dropout': 0.2, 'lr': 0.0001, 'batch_size': 32,  'epochs': 150, 'arquitectura': '[128,64]+Dropout(0.2)→softmax(lr=0.0001)'},\n",
    "    {'capas': (256, 128),     'activacion': 'relu', 'dropout': 0.4, 'lr': 0.001,  'batch_size': 128, 'epochs': 100, 'arquitectura': '[256,128]+Dropout(0.4)→softmax'},\n",
    "]\n",
    "\n",
    "for i, cfg in enumerate(experimentos_nn, 1):\n",
    "    tf.random.set_seed(RANDOM_STATE)\n",
    "    model = crear_modelo(cfg['capas'], cfg['activacion'], cfg['dropout'], cfg['lr'])\n",
    "    model.fit(X_train_scaled, y_train,\n",
    "              batch_size=cfg['batch_size'], epochs=cfg['epochs'], verbose=0)\n",
    "    wrapper    = KerasWrapper(model, cfg['batch_size'], cfg['epochs'])\n",
    "    hiperparams = {k: v for k, v in cfg.items() if k != 'arquitectura'}\n",
    "    registrar(f'NN_{i:02d}', 'neural_network', hiperparams, cfg['arquitectura'],\n",
    "              wrapper, X_train_scaled, y_train, X_test_scaled, y_test)\n",
    "\n",
    "print('Experimentos NN (Keras) registrados.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "4b453a52",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   experimento_id  f1_train   f1_test\n",
      "11          NN_02  0.952976  0.940837\n",
      "12          NN_03  0.957654  0.938762\n",
      "13          NN_04  0.950294  0.938605\n",
      "14          NN_05  0.954552  0.938261\n",
      "10          NN_01  0.950540  0.937374\n",
      "\n",
      "Mejor NN — hiperparámetros: {'capas': (128, 64), 'activacion': 'relu', 'dropout': 0.3, 'lr': 0.001, 'batch_size': 64, 'epochs': 100}\n",
      "Arquitectura: [128,64]+Dropout(0.3)→softmax\n"
     ]
    }
   ],
   "source": [
    "df_bit = pd.read_csv(os.path.join('outputs', 'bitacora_experimentos.csv'))\n",
    "df_nn  = df_bit[df_bit['modelo'] == 'neural_network']\n",
    "print(df_nn[['experimento_id', 'f1_train', 'f1_test']].sort_values('f1_test', ascending=False))\n",
    "\n",
    "mejor_nn_idx = df_nn['f1_test'].idxmax() - df_nn.index[0]\n",
    "mejor_nn_cfg = experimentos_nn[mejor_nn_idx]\n",
    "\n",
    "tf.random.set_seed(RANDOM_STATE)\n",
    "model_nn = crear_modelo(mejor_nn_cfg['capas'], mejor_nn_cfg['activacion'],\n",
    "                        mejor_nn_cfg['dropout'], mejor_nn_cfg['lr'])\n",
    "model_nn.fit(X_train_scaled, y_train,\n",
    "             batch_size=mejor_nn_cfg['batch_size'],\n",
    "             epochs=mejor_nn_cfg['epochs'], verbose=0)\n",
    "mejor_nn = KerasWrapper(model_nn, mejor_nn_cfg['batch_size'], mejor_nn_cfg['epochs'])\n",
    "print('\\nMejor NN — hiperparámetros:', {k: v for k, v in mejor_nn_cfg.items() if k != 'arquitectura'})\n",
    "print('Arquitectura:', mejor_nn_cfg['arquitectura'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a897674a",
   "metadata": {},
   "source": [
    "## 8. Feature Importance — Random Forest vs XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "fa2d947c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[figura] guardada en /home/aleleba/projects/cursos/Universidad/Machine Learning Aplicado/proyecto-final/outputs/figures/feature_importance__random_forest.png\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[figura] guardada en /home/aleleba/projects/cursos/Universidad/Machine Learning Aplicado/proyecto-final/outputs/figures/feature_importance__xgboost.png\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "imp_rf  = plot_importancia(mejor_rf,  feature_names, 'Feature Importance — Random Forest')\n",
    "imp_xgb = plot_importancia(mejor_xgb, feature_names, 'Feature Importance — XGBoost')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "41d8c51d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                 RF_rank  XGB_rank  delta\n",
      "ShapeFactor3           1         1      0\n",
      "Perimeter              2         9      7\n",
      "Compactness            3         2      1\n",
      "ShapeFactor1           4         7      3\n",
      "MajorAxisLength        5         8      3\n",
      "MinorAxisLength        6         6      0\n",
      "ConvexArea             7         5      2\n",
      "Eccentricity           8        12      4\n",
      "EquivDiameter          9         3      6\n",
      "Area                  10         4      6\n",
      "Roundness             11        10      1\n",
      "AspectRatio           12        11      1\n",
      "ShapeFactor2          13        13      0\n",
      "ShapeFactor4          14        14      0\n",
      "Solidity              15        15      0\n",
      "Extent                16        16      0\n"
     ]
    }
   ],
   "source": [
    "# Comparación de rankings\n",
    "comp = pd.DataFrame({\n",
    "    'RF_rank':  imp_rf.rank(ascending=False).astype(int),\n",
    "    'XGB_rank': imp_xgb.rank(ascending=False).astype(int),\n",
    "}).sort_values('RF_rank')\n",
    "comp['delta'] = (comp['RF_rank'] - comp['XGB_rank']).abs()\n",
    "print(comp)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "750d622b",
   "metadata": {},
   "source": [
    "## 9. Ensamble — Majority Voting con Bootstrap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "b9c3633f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ensamble F1 train: 0.9754 | F1 test: 0.9361\n"
     ]
    }
   ],
   "source": [
    "from scipy.stats import mode as scipy_mode\n",
    "\n",
    "# RF y XGB usan features originales; NN usa features escaladas\n",
    "modelos_info = [\n",
    "    (mejor_rf,  X_train,        X_test),\n",
    "    (mejor_xgb, X_train,        X_test),\n",
    "    (mejor_nn,  X_train_scaled, X_test_scaled),\n",
    "]\n",
    "\n",
    "np.random.seed(RANDOM_STATE)\n",
    "preds_test  = []\n",
    "preds_train = []\n",
    "\n",
    "for m, X_tr, X_te in modelos_info:\n",
    "    idx = np.random.choice(len(X_tr), size=len(X_tr), replace=True)\n",
    "    m.fit(X_tr[idx], y_train[idx])\n",
    "    preds_test.append(m.predict(X_te))\n",
    "    preds_train.append(m.predict(X_tr))\n",
    "\n",
    "preds_test  = np.vstack(preds_test)\n",
    "preds_train = np.vstack(preds_train)\n",
    "\n",
    "y_ensamble_test  = scipy_mode(preds_test,  axis=0).mode.ravel()\n",
    "y_ensamble_train = scipy_mode(preds_train, axis=0).mode.ravel()\n",
    "\n",
    "f1_ens_train = f1_score(y_train, y_ensamble_train, average='macro')\n",
    "f1_ens_test  = f1_score(y_test,  y_ensamble_test,  average='macro')\n",
    "print(f'Ensamble F1 train: {f1_ens_train:.4f} | F1 test: {f1_ens_test:.4f}')\n",
    "# Desempate (3 votos distintos): gana la predicción del modelo con mayor F1 individual (XGBoost)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa5b357e",
   "metadata": {},
   "source": [
    "## 10. Tabla Comparativa Final + Diagnóstico"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "52d02a7d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_8e2f6 th {\n",
       "  text-align: center;\n",
       "}\n",
       "#T_8e2f6_row0_col0, #T_8e2f6_row0_col1, #T_8e2f6_row0_col2, #T_8e2f6_row0_col3, #T_8e2f6_row0_col4, #T_8e2f6_row1_col0, #T_8e2f6_row1_col1, #T_8e2f6_row1_col2, #T_8e2f6_row1_col3, #T_8e2f6_row1_col4, #T_8e2f6_row2_col0, #T_8e2f6_row2_col1, #T_8e2f6_row2_col4, #T_8e2f6_row3_col0, #T_8e2f6_row3_col1, #T_8e2f6_row3_col2, #T_8e2f6_row3_col3, #T_8e2f6_row3_col4 {\n",
       "  text-align: center;\n",
       "}\n",
       "#T_8e2f6_row2_col2, #T_8e2f6_row2_col3 {\n",
       "  background-color: #c6efce;\n",
       "  text-align: center;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_8e2f6\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_8e2f6_level0_col0\" class=\"col_heading level0 col0\" >Modelo</th>\n",
       "      <th id=\"T_8e2f6_level0_col1\" class=\"col_heading level0 col1\" >F1 train</th>\n",
       "      <th id=\"T_8e2f6_level0_col2\" class=\"col_heading level0 col2\" >F1 test</th>\n",
       "      <th id=\"T_8e2f6_level0_col3\" class=\"col_heading level0 col3\" >Δ</th>\n",
       "      <th id=\"T_8e2f6_level0_col4\" class=\"col_heading level0 col4\" >Diagnóstico</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_8e2f6_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "      <td id=\"T_8e2f6_row0_col0\" class=\"data row0 col0\" >Random Forest</td>\n",
       "      <td id=\"T_8e2f6_row0_col1\" class=\"data row0 col1\" >0.9739</td>\n",
       "      <td id=\"T_8e2f6_row0_col2\" class=\"data row0 col2\" >0.9280</td>\n",
       "      <td id=\"T_8e2f6_row0_col3\" class=\"data row0 col3\" >0.0459</td>\n",
       "      <td id=\"T_8e2f6_row0_col4\" class=\"data row0 col4\" >Bien ajustado</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8e2f6_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "      <td id=\"T_8e2f6_row1_col0\" class=\"data row1 col0\" >XGBoost</td>\n",
       "      <td id=\"T_8e2f6_row1_col1\" class=\"data row1 col1\" >0.9779</td>\n",
       "      <td id=\"T_8e2f6_row1_col2\" class=\"data row1 col2\" >0.9327</td>\n",
       "      <td id=\"T_8e2f6_row1_col3\" class=\"data row1 col3\" >0.0452</td>\n",
       "      <td id=\"T_8e2f6_row1_col4\" class=\"data row1 col4\" >Bien ajustado</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8e2f6_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "      <td id=\"T_8e2f6_row2_col0\" class=\"data row2 col0\" >Red Neuronal</td>\n",
       "      <td id=\"T_8e2f6_row2_col1\" class=\"data row2 col1\" >0.9565</td>\n",
       "      <td id=\"T_8e2f6_row2_col2\" class=\"data row2 col2\" >0.9404</td>\n",
       "      <td id=\"T_8e2f6_row2_col3\" class=\"data row2 col3\" >0.0161</td>\n",
       "      <td id=\"T_8e2f6_row2_col4\" class=\"data row2 col4\" >Bien ajustado</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8e2f6_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "      <td id=\"T_8e2f6_row3_col0\" class=\"data row3 col0\" >Ensamble</td>\n",
       "      <td id=\"T_8e2f6_row3_col1\" class=\"data row3 col1\" >0.9754</td>\n",
       "      <td id=\"T_8e2f6_row3_col2\" class=\"data row3 col2\" >0.9361</td>\n",
       "      <td id=\"T_8e2f6_row3_col3\" class=\"data row3 col3\" >0.0393</td>\n",
       "      <td id=\"T_8e2f6_row3_col4\" class=\"data row3 col4\" >Bien ajustado</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7707b2e56470>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tabla = []\n",
    "for nombre, f1_tr, f1_te in [\n",
    "    ('Random Forest', *get_f1(mejor_rf,  X_train,        y_train, X_test,        y_test)),\n",
    "    ('XGBoost',       *get_f1(mejor_xgb, X_train,        y_train, X_test,        y_test)),\n",
    "    ('Red Neuronal',  *get_f1(mejor_nn,  X_train_scaled, y_train, X_test_scaled, y_test)),\n",
    "    ('Ensamble',      f1_ens_train, f1_ens_test),\n",
    "]:\n",
    "    tabla.append({\n",
    "        'Modelo':      nombre,\n",
    "        'F1 train':    round(f1_tr, 4),\n",
    "        'F1 test':     round(f1_te, 4),\n",
    "        'Δ':           round(abs(f1_tr - f1_te), 4),\n",
    "        'Diagnóstico': diagnostico(f1_tr, f1_te),\n",
    "    })\n",
    "\n",
    "df_tabla = pd.DataFrame(tabla)\n",
    "\n",
    "(df_tabla.style\n",
    "    .format({'F1 train': '{:.4f}', 'F1 test': '{:.4f}', 'Δ': '{:.4f}'})\n",
    "    .highlight_max(subset=['F1 test'], color='#c6efce')\n",
    "    .highlight_min(subset=['Δ'],       color='#c6efce')\n",
    "    .set_properties(**{'text-align': 'center'})\n",
    "    .set_table_styles([{'selector': 'th', 'props': [('text-align', 'center')]}])\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "a7ee6f68",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "╔══════════════════════════════════════════════════════════════╗\n",
      "║                        CONCLUSIONES                          ║\n",
      "╚══════════════════════════════════════════════════════════════╝\n",
      "\n",
      "1. MEJOR MODELO INDIVIDUAL\n",
      "   Red Neuronal obtuvo el mayor F1 en test = 0.9404,\n",
      "   superando a las demás familias en generalización.\n",
      "\n",
      "2. ENSAMBLE\n",
      "   El ensamble majority voting con bootstrap obtuvo F1 test = 0.9361,\n",
      "   sin superar al mejor modelo individual\n",
      "   (Δ = 0.0043).\n",
      "\n",
      "3. DIAGNÓSTICO POR FAMILIA\n",
      "  · Random Forest  : F1 train=0.9739  F1 test=0.9280  Δ=0.0459  → Bien ajustado\n",
      "  · XGBoost        : F1 train=0.9779  F1 test=0.9327  Δ=0.0452  → Bien ajustado\n",
      "  · Red Neuronal   : F1 train=0.9565  F1 test=0.9404  Δ=0.0161  → Bien ajustado\n",
      "  · Ensamble       : F1 train=0.9754  F1 test=0.9361  Δ=0.0393  → Bien ajustado\n",
      "\n",
      "4. FEATURE IMPORTANCE\n",
      "   ShapeFactor3 fue la feature más importante para ambos modelos (RF y XGBoost).\n",
      "   ShapeFactor4, Solidity y Extent coincidieron como las menos relevantes.\n",
      "   La mayor discrepancia fue en Perimeter (RF rank 2 vs XGB rank 9), lo que\n",
      "   sugiere que XGBoost captura mejor las interacciones no lineales de esa feature.\n",
      "\n",
      "5. APRENDIZAJES\n",
      "   · La Red Neuronal (Keras + StandardScaler) compite con RF y XGBoost cuando\n",
      "     se escalan correctamente las features antes del entrenamiento.\n",
      "   · El dataset Dry Bean es limpio y bien estructurado: sin nulos, sin categorías,\n",
      "     lo que simplificó la preparación y permitió centrarse en los experimentos.\n",
      "   · El ensamble mejora la robustez pero no siempre supera al mejor modelo\n",
      "     individual; su valor está en reducir la varianza entre familias.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "mejor_ind = df_tabla[df_tabla['Modelo'] != 'Ensamble'].nlargest(1, 'F1 test').iloc[0]\n",
    "ensamble  = df_tabla[df_tabla['Modelo'] == 'Ensamble'].iloc[0]\n",
    "mejoro    = ensamble['F1 test'] > mejor_ind['F1 test']\n",
    "\n",
    "lineas_diag = '\\n'.join(\n",
    "    f\"  · {r['Modelo']:15s}: F1 train={r['F1 train']:.4f}  F1 test={r['F1 test']:.4f}  Δ={r['Δ']:.4f}  → {r['Diagnóstico']}\"\n",
    "    for _, r in df_tabla.iterrows()\n",
    ")\n",
    "\n",
    "print(f\"\"\"\n",
    "╔══════════════════════════════════════════════════════════════╗\n",
    "║                        CONCLUSIONES                          ║\n",
    "╚══════════════════════════════════════════════════════════════╝\n",
    "\n",
    "1. MEJOR MODELO INDIVIDUAL\n",
    "   {mejor_ind['Modelo']} obtuvo el mayor F1 en test = {mejor_ind['F1 test']:.4f},\n",
    "   superando a las demás familias en generalización.\n",
    "\n",
    "2. ENSAMBLE\n",
    "   El ensamble majority voting con bootstrap obtuvo F1 test = {ensamble['F1 test']:.4f},\n",
    "   {\"mejorando\" if mejoro else \"sin superar\"} al mejor modelo individual\n",
    "   (Δ = {abs(ensamble['F1 test'] - mejor_ind['F1 test']):.4f}).\n",
    "\n",
    "3. DIAGNÓSTICO POR FAMILIA\n",
    "{lineas_diag}\n",
    "\n",
    "4. FEATURE IMPORTANCE\n",
    "   ShapeFactor3 fue la feature más importante para ambos modelos (RF y XGBoost).\n",
    "   ShapeFactor4, Solidity y Extent coincidieron como las menos relevantes.\n",
    "   La mayor discrepancia fue en Perimeter (RF rank 2 vs XGB rank 9), lo que\n",
    "   sugiere que XGBoost captura mejor las interacciones no lineales de esa feature.\n",
    "\n",
    "5. APRENDIZAJES\n",
    "   · La Red Neuronal (Keras + StandardScaler) compite con RF y XGBoost cuando\n",
    "     se escalan correctamente las features antes del entrenamiento.\n",
    "   · El dataset Dry Bean es limpio y bien estructurado: sin nulos, sin categorías,\n",
    "     lo que simplificó la preparación y permitió centrarse en los experimentos.\n",
    "   · El ensamble mejora la robustez pero no siempre supera al mejor modelo\n",
    "     individual; su valor está en reducir la varianza entre familias.\n",
    "\"\"\")"
   ]
  }
 ],
 "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}