notebooks/ensemble_ex_04.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a9da1f38",
   "metadata": {},
   "source": [
    "# 📝 Exercise M6.04\n",
    "\n",
    "The aim of this exercise is to:\n",
    "\n",
    "* verify if a GBDT tends to overfit if the number of estimators is not\n",
    "  appropriate as previously seen for AdaBoost;\n",
    "* use the early-stopping strategy to avoid adding unnecessary trees, to\n",
    "  get the best statistical performances.\n",
    "\n",
    "We will use the California housing dataset to conduct our experiments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "85259f15",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-06-21T11:45:33.644038Z",
     "start_time": "2021-06-21T11:45:33.247974Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_california_housing\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "data, target = fetch_california_housing(return_X_y=True, as_frame=True)\n",
    "target *= 100  # rescale the target in k$\n",
    "data_train, data_test, target_train, target_test = train_test_split(\n",
    "    data, target, random_state=0, test_size=0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a04a2380",
   "metadata": {},
   "source": [
    "<div class=\"admonition note alert alert-info\">\n",
    "<p class=\"first admonition-title\" style=\"font-weight: bold;\">Note</p>\n",
    "<p class=\"last\">If you want a deeper overview regarding this dataset, you can refer to the\n",
    "Appendix - Datasets description section at the end of this MOOC.</p>\n",
    "</div>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10e41c5e",
   "metadata": {},
   "source": [
    "Similarly to the previous exercise, create a gradient boosting decision tree\n",
    "and create a validation curve to assess the impact of the number of trees\n",
    "on the statistical performance of the model. Use the mean absolute error\n",
    "to assess the statistical performance of the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "36f9826a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-06-21T11:48:03.591402Z",
     "start_time": "2021-06-21T11:46:09.874533Z"
    }
   },
   "outputs": [],
   "source": [
    "# Write your code here.\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.model_selection import validation_curve\n",
    "import numpy as np\n",
    "# nb_estimators = np.unique(np.logspace(0, 1.8, num=30).astype(int))\n",
    "\n",
    "nb_estimators = list(range(1, 61))\n",
    "regressor= GradientBoostingRegressor()\n",
    "\n",
    "train_scores, test_scores = validation_curve(\n",
    "    regressor, data_train, target_train, param_name=\"n_estimators\", param_range=nb_estimators,\n",
    "    scoring=\"neg_mean_absolute_error\")\n",
    "train_errors, test_errors = -train_scores, -test_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b18855c0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-06-21T11:48:03.868121Z",
     "start_time": "2021-06-21T11:48:03.592373Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.errorbar(nb_estimators, train_errors.mean(axis=1),\n",
    "             yerr=train_errors.std(axis=1), label=\"Training score\",\n",
    "             alpha=0.7)\n",
    "plt.errorbar(nb_estimators, test_errors.mean(axis=1),\n",
    "             yerr=test_errors.std(axis=1), label=\"Cross-validation score\",\n",
    "             alpha=0.7)\n",
    "\n",
    "plt.legend()\n",
    "plt.ylabel(\"Mean absolute error in k$\\n(smaller is better)\")\n",
    "plt.xlabel(\"# estimators\")\n",
    "_ = plt.title(\"Validation curve for GradientBoosting regressor\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f669351",
   "metadata": {},
   "source": [
    "Unlike AdaBoost, the gradient boosting model will always improve when\n",
    "increasing the number of trees in the ensemble. However, it will reach a\n",
    "plateau where adding new trees will just make fitting and scoring slower.\n",
    "\n",
    "To avoid adding new unnecessary tree, gradient boosting offers an\n",
    "early-stopping option. Internally, the algorithm will use an out-of-sample\n",
    "set to compute the statistical performance of the model at each addition of a\n",
    "tree. Thus, if the statistical performance are not improving for several\n",
    "iterations, it will stop adding trees.\n",
    "\n",
    "Now, create a gradient-boosting model with `n_estimators=1000`. This number\n",
    "of trees will be too large. Change the parameter `n_iter_no_change` such\n",
    "that the gradient boosting fitting will stop after adding 5 trees that do not\n",
    "improve the overall statistical performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "2d78b131",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-06-21T11:50:30.329995Z",
     "start_time": "2021-06-21T11:50:28.682123Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingRegressor(n_estimators=1000, n_iter_no_change=5)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Write your code here.\n",
    "# Write your code here.\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.model_selection import validation_curve\n",
    "import numpy as np\n",
    "# nb_estimators = np.unique(np.logspace(0, 1.8, num=30).astype(int))\n",
    "\n",
    "regressor= GradientBoostingRegressor(n_estimators=1000, n_iter_no_change=5)\n",
    "\n",
    "regressor.fit(data_train, target_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "5a4a8fd2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-06-21T11:52:13.536522Z",
     "start_time": "2021-06-21T11:52:13.530093Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([11793.89685785, 10685.26616378,  9725.07767252,  8930.38072385,\n",
       "        8289.47151565,  7743.72149217,  7294.13141644,  6873.9459535 ,\n",
       "        6549.72811261,  6249.18745483,  5984.91328524,  5760.54008053,\n",
       "        5561.71088551,  5388.96434849,  5230.28446401,  5079.5574533 ,\n",
       "        4944.61541394,  4825.8596265 ,  4701.64944177,  4576.00889395,\n",
       "        4484.90736999,  4387.64315481,  4225.29951863,  4092.25588248,\n",
       "        4024.53753004,  3912.51177803,  3855.93827473,  3807.64304535,\n",
       "        3754.44396364,  3703.48236891,  3661.90177424,  3591.09927787,\n",
       "        3554.23628572,  3488.04735895,  3446.64885269,  3417.60550963,\n",
       "        3384.66217236,  3316.92853451,  3296.11077929,  3269.85861774,\n",
       "        3245.58822971,  3219.40570996,  3195.49482891,  3167.63438542,\n",
       "        3135.31470114,  3114.24368482,  3098.1824098 ,  3083.95235398,\n",
       "        3048.35908146,  3031.61960344,  3002.96127233,  2990.48467556,\n",
       "        2977.1600505 ,  2962.91095329,  2944.35221016,  2929.7177072 ,\n",
       "        2911.68892384,  2901.57597471,  2892.6838571 ,  2882.87208554,\n",
       "        2872.71300531,  2857.68712253,  2826.84156517,  2815.35626231,\n",
       "        2790.17922856,  2783.60046764,  2774.46192413,  2765.40476373,\n",
       "        2758.41878593,  2751.63350884,  2742.77332595,  2736.52823541,\n",
       "        2728.64857746,  2721.5928189 ,  2714.26212657,  2697.68890457,\n",
       "        2688.5201449 ,  2683.57719289,  2669.98027635,  2664.38045809,\n",
       "        2657.14948247,  2650.81848205,  2646.88811034,  2641.94821385,\n",
       "        2629.27343187,  2625.0437259 ,  2614.33157828,  2597.90787888,\n",
       "        2589.69813152,  2579.23857051,  2573.70771791,  2553.87585318,\n",
       "        2540.30318735,  2534.50098415,  2526.46832136,  2523.34559516,\n",
       "        2517.58019529,  2511.37429314,  2505.09469358,  2499.38613657,\n",
       "        2491.17541596,  2487.72638239,  2482.53050518,  2466.63079566,\n",
       "        2455.9977533 ,  2447.65475896,  2442.9391327 ,  2437.98708354,\n",
       "        2434.85494092,  2432.03160311,  2427.17512993,  2423.5330482 ,\n",
       "        2413.84699267,  2405.87248711,  2392.96521056,  2385.99434717,\n",
       "        2380.37098589,  2377.58214276,  2376.35698924,  2368.49659702,\n",
       "        2356.06222467,  2346.38918623,  2335.55385778,  2328.23735228,\n",
       "        2322.83604505,  2317.97416818,  2316.13481158,  2310.54884816,\n",
       "        2307.13399325])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regressor.train_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "683fa76e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-06-21T11:53:25.274704Z",
     "start_time": "2021-06-21T11:53:25.265473Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "129"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regressor.n_estimators_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7f7cf217",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "jupytext": {
   "cell_metadata_filter": "-all",
   "main_language": "python",
   "notebook_metadata_filter": "-all"
  },
  "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": "3.9.4"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}