mlessentials/Lab08/Examples/tuning_using_gridsearchcv.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "aBtvWJNQhBM5"
   },
   "outputs": [],
   "source": [
    "from sklearn import model_selection, datasets, neighbors\n",
    "\n",
    "# load the data\n",
    "cancer = datasets.load_breast_cancer()\n",
    "\n",
    "# target\n",
    "y = cancer.target\n",
    "\n",
    "# features\n",
    "X = cancer.data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "p6fQASEChP03"
   },
   "outputs": [],
   "source": [
    "# initalize the estimator\n",
    "knn = neighbors.KNeighborsClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "cADtTvf6hb3O"
   },
   "outputs": [],
   "source": [
    "# grid contains k and the weight function\n",
    "grid = {\n",
    "    'n_neighbors': [1, 3, 5, 7],\n",
    "    'weights': ['uniform', 'distance']\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZZ_uK5A0ho67"
   },
   "outputs": [],
   "source": [
    "# set up the grid search with scoring on precsions and number of folds = 10\n",
    "gscv = model_selection.GridSearchCV(estimator=knn, param_grid=grid, scoring='precision', cv=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 263
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1591,
     "status": "ok",
     "timestamp": 1571314594262,
     "user": {
      "displayName": "Andrew Worsley",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mAp-Td-yKvu76Tg0Swzal8U17btuwNIXFmWVwZo=s64",
      "userId": "11337101975325054847"
     },
     "user_tz": -660
    },
    "id": "tLqJ_3fBh2Bh",
    "outputId": "bf075dfd-0a91-4171-951e-6735b6885c98"
   },
   "outputs": [],
   "source": [
    "# start the search\n",
    "gscv.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 454
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1583,
     "status": "ok",
     "timestamp": 1571314594263,
     "user": {
      "displayName": "Andrew Worsley",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mAp-Td-yKvu76Tg0Swzal8U17btuwNIXFmWVwZo=s64",
      "userId": "11337101975325054847"
     },
     "user_tz": -660
    },
    "id": "yPZkBBZGiWuC",
    "outputId": "ae372db9-8dee-4a68-a564-48dc61d693b8"
   },
   "outputs": [],
   "source": [
    "# view the results\n",
    "print(gscv.cv_results_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 121
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1576,
     "status": "ok",
     "timestamp": 1571314594265,
     "user": {
      "displayName": "Andrew Worsley",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mAp-Td-yKvu76Tg0Swzal8U17btuwNIXFmWVwZo=s64",
      "userId": "11337101975325054847"
     },
     "user_tz": -660
    },
    "id": "H6n1RzMiifkZ",
    "outputId": "cb139684-1c11-4e63-827a-0d8b2dd496e7"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "# convert the results dictionary to a dataframe\n",
    "results = pd.DataFrame(gscv.cv_results_)\n",
    "\n",
    "# select just the hyperparameterizations tried, the mean test scores, order by score and show the top 5 models\n",
    "print(\n",
    "results.loc[:,['params','mean_test_score']].sort_values('mean_test_score', ascending=False).head(5)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1323,
     "status": "ok",
     "timestamp": 1571314633729,
     "user": {
      "displayName": "Andrew Worsley",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AAuE7mAp-Td-yKvu76Tg0Swzal8U17btuwNIXFmWVwZo=s64",
      "userId": "11337101975325054847"
     },
     "user_tz": -660
    },
    "id": "I92n0QuvpM03",
    "outputId": "28021da8-996b-4b0b-c821-8ac571492b1b"
   },
   "outputs": [],
   "source": [
    "# visualize the result\n",
    "results.loc[:,['params','mean_test_score']].plot.barh(x = 'params')"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "tuning_using_gridsearchcv.ipynb",
   "provenance": []
  },
  "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": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}