mlessentials/Lab08/Exercise8.03/Excerise_8_3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1482,
     "status": "ok",
     "timestamp": 1596620773872,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "jC-tIAMXIT_l"
   },
   "outputs": [],
   "source": [
    "from sklearn import datasets\n",
    "\n",
    "# import data\n",
    "digits = datasets.load_digits()\n",
    "\n",
    "# target\n",
    "y = digits.target\n",
    "\n",
    "# features\n",
    "X = digits.data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1767,
     "status": "ok",
     "timestamp": 1596620863966,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "IyLRPz2EIWJ_"
   },
   "outputs": [],
   "source": [
    "from sklearn import ensemble\n",
    "\n",
    "# an ensemble of 100 estimators\n",
    "rfc = ensemble.RandomForestClassifier(n_estimators=100, random_state=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1245,
     "status": "ok",
     "timestamp": 1596620863968,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "Tr0ZLj2yJnM2",
    "outputId": "fc20432b-d520-4462-c8b6-5b3cd950a17f"
   },
   "outputs": [],
   "source": [
    "# how many features do we have in our dataset?\n",
    "n_features = X.shape[1]\n",
    "\n",
    "print(n_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 911,
     "status": "ok",
     "timestamp": 1596620864645,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "vuJQ0pJuIvCL"
   },
   "outputs": [],
   "source": [
    "from scipy import stats\n",
    "import numpy as np\n",
    "np.random.seed(100)\n",
    "\n",
    "# we would like to smaple from criterion and max_features as discrete uniform distributions\n",
    "param_dist = {\n",
    "    'criterion': ['gini', 'entropy'],\n",
    "    'max_features': stats.randint(low=1, high=n_features)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1096,
     "status": "ok",
     "timestamp": 1596620866276,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "3-vP3LMpKQ2S"
   },
   "outputs": [],
   "source": [
    "from sklearn import model_selection\n",
    "\n",
    "# setting up the random search sampling 50 times and conducting 5-fold cross-validation\n",
    "rscv = model_selection.RandomizedSearchCV(estimator=rfc, param_distributions=param_dist, n_iter=50, cv=5, scoring='accuracy', random_state=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 444
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 234185,
     "status": "ok",
     "timestamp": 1596621100375,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "yOporT_hKaTd",
    "outputId": "d34916b6-03da-4183-dd64-904177f2f0ff"
   },
   "outputs": [],
   "source": [
    "# start the process\n",
    "rscv.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 124
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1908,
     "status": "ok",
     "timestamp": 1596621454897,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "Hn06r3ktKkln",
    "outputId": "2cdb8c61-2f87-4fb8-bac6-accbac0e5166"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "# convert the dictionary of results to a pandas dataframe\n",
    "results = pd.DataFrame(rscv.cv_results_)\n",
    "\n",
    "# removing duplication\n",
    "distinct_results = results.loc[:,['params','mean_test_score']]\n",
    "\n",
    "# convert the params dictionaries to string data types\n",
    "distinct_results.loc[:,'params'] = distinct_results.loc[:,'params'].astype('str')\n",
    "\n",
    "# remove duplicates\n",
    "distinct_results.drop_duplicates(inplace=True)\n",
    "\n",
    "# look at the top 5 best hyperparamaterizations\n",
    "print(distinct_results.sort_values('mean_test_score', ascending=False).head(5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 283
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1979,
     "status": "ok",
     "timestamp": 1596621460667,
     "user": {
      "displayName": "Mahesh Dhyani",
      "photoUrl": "",
      "userId": "02713575687182950579"
     },
     "user_tz": -330
    },
    "id": "zctH43A48Z0r",
    "outputId": "9de9e641-e20b-460e-8336-3c9d528c8726"
   },
   "outputs": [],
   "source": [
    "# top performing models\n",
    "distinct_results[distinct_results.mean_test_score > 0.93].sort_values('mean_test_score').plot.barh(x='params', xlim=(0.9))"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "Excerise_8_3.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
}