Interpretable Classification Methods.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup a classification experiment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "df = pd.read_csv(\n",
    "    \"https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data\",\n",
    "    header=None)\n",
    "df.columns = [\n",
    "    \"Age\", \"WorkClass\", \"fnlwgt\", \"Education\", \"EducationNum\",\n",
    "    \"MaritalStatus\", \"Occupation\", \"Relationship\", \"Race\", \"Gender\",\n",
    "    \"CapitalGain\", \"CapitalLoss\", \"HoursPerWeek\", \"NativeCountry\", \"Income\"\n",
    "]\n",
    "# df = df.sample(frac=0.1, random_state=1)\n",
    "train_cols = df.columns[0:-1]\n",
    "label = df.columns[-1]\n",
    "X = df[train_cols]\n",
    "y = df[label].apply(lambda x: 0 if x == \" <=50K\" else 1) #Turning response into 0 and 1\n",
    "\n",
    "seed = 1\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Age</th>\n",
       "      <th>WorkClass</th>\n",
       "      <th>fnlwgt</th>\n",
       "      <th>Education</th>\n",
       "      <th>EducationNum</th>\n",
       "      <th>MaritalStatus</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>Relationship</th>\n",
       "      <th>Race</th>\n",
       "      <th>Gender</th>\n",
       "      <th>CapitalGain</th>\n",
       "      <th>CapitalLoss</th>\n",
       "      <th>HoursPerWeek</th>\n",
       "      <th>NativeCountry</th>\n",
       "      <th>Income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>77516</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>2174</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>50</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>83311</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>215646</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>53</td>\n",
       "      <td>Private</td>\n",
       "      <td>234721</td>\n",
       "      <td>11th</td>\n",
       "      <td>7</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>338409</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Wife</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>Cuba</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32556</td>\n",
       "      <td>27</td>\n",
       "      <td>Private</td>\n",
       "      <td>257302</td>\n",
       "      <td>Assoc-acdm</td>\n",
       "      <td>12</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Tech-support</td>\n",
       "      <td>Wife</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>38</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32557</td>\n",
       "      <td>40</td>\n",
       "      <td>Private</td>\n",
       "      <td>154374</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32558</td>\n",
       "      <td>58</td>\n",
       "      <td>Private</td>\n",
       "      <td>151910</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Widowed</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Unmarried</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32559</td>\n",
       "      <td>22</td>\n",
       "      <td>Private</td>\n",
       "      <td>201490</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32560</td>\n",
       "      <td>52</td>\n",
       "      <td>Self-emp-inc</td>\n",
       "      <td>287927</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Wife</td>\n",
       "      <td>White</td>\n",
       "      <td>Female</td>\n",
       "      <td>15024</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>32561 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Age          WorkClass  fnlwgt    Education  EducationNum  \\\n",
       "0       39          State-gov   77516    Bachelors            13   \n",
       "1       50   Self-emp-not-inc   83311    Bachelors            13   \n",
       "2       38            Private  215646      HS-grad             9   \n",
       "3       53            Private  234721         11th             7   \n",
       "4       28            Private  338409    Bachelors            13   \n",
       "...    ...                ...     ...          ...           ...   \n",
       "32556   27            Private  257302   Assoc-acdm            12   \n",
       "32557   40            Private  154374      HS-grad             9   \n",
       "32558   58            Private  151910      HS-grad             9   \n",
       "32559   22            Private  201490      HS-grad             9   \n",
       "32560   52       Self-emp-inc  287927      HS-grad             9   \n",
       "\n",
       "             MaritalStatus          Occupation    Relationship    Race  \\\n",
       "0            Never-married        Adm-clerical   Not-in-family   White   \n",
       "1       Married-civ-spouse     Exec-managerial         Husband   White   \n",
       "2                 Divorced   Handlers-cleaners   Not-in-family   White   \n",
       "3       Married-civ-spouse   Handlers-cleaners         Husband   Black   \n",
       "4       Married-civ-spouse      Prof-specialty            Wife   Black   \n",
       "...                    ...                 ...             ...     ...   \n",
       "32556   Married-civ-spouse        Tech-support            Wife   White   \n",
       "32557   Married-civ-spouse   Machine-op-inspct         Husband   White   \n",
       "32558              Widowed        Adm-clerical       Unmarried   White   \n",
       "32559        Never-married        Adm-clerical       Own-child   White   \n",
       "32560   Married-civ-spouse     Exec-managerial            Wife   White   \n",
       "\n",
       "        Gender  CapitalGain  CapitalLoss  HoursPerWeek   NativeCountry  Income  \n",
       "0         Male         2174            0            40   United-States   <=50K  \n",
       "1         Male            0            0            13   United-States   <=50K  \n",
       "2         Male            0            0            40   United-States   <=50K  \n",
       "3         Male            0            0            40   United-States   <=50K  \n",
       "4       Female            0            0            40            Cuba   <=50K  \n",
       "...        ...          ...          ...           ...             ...     ...  \n",
       "32556   Female            0            0            38   United-States   <=50K  \n",
       "32557     Male            0            0            40   United-States    >50K  \n",
       "32558   Female            0            0            40   United-States   <=50K  \n",
       "32559     Male            0            0            20   United-States   <=50K  \n",
       "32560   Female        15024            0            40   United-States    >50K  \n",
       "\n",
       "[32561 rows x 15 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Explore the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631271193232/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631271193232/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from interpret import show\n",
    "from interpret.data import ClassHistogram\n",
    "\n",
    "hist = ClassHistogram().explain_data(X_train, y_train, name = 'Train Data')\n",
    "show(hist)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train the Explainable Boosting Machine (EBM)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ExplainableBoostingClassifier(binning_strategy='uniform',\n",
       "               data_n_episodes=2000, early_stopping_run_length=50,\n",
       "               early_stopping_tolerance=1e-05,\n",
       "               feature_names=['Age', 'WorkClass', 'fnlwgt', 'Education', 'EducationNum', 'MaritalStatus', 'Occupation', 'Relationship', 'Race', 'Gender', 'CapitalGain', 'CapitalLoss', 'HoursPerWeek', 'NativeCountry'],\n",
       "               feature_step_n_inner_bags=0,\n",
       "               feature_types=['continuous', 'categorical', 'continuous', 'categorical', 'continuous', 'categorical', 'categorical', 'categorical', 'categorical', 'categorical', 'continuous', 'continuous', 'continuous', 'categorical'],\n",
       "               holdout_size=0.15, holdout_split=0.15, interactions=0,\n",
       "               learning_rate=0.01, max_tree_splits=2,\n",
       "               min_cases_for_splits=2, n_estimators=16, n_jobs=-2,\n",
       "               random_state=1, schema=None, scoring=None,\n",
       "               training_step_episodes=1)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from interpret.glassbox import ExplainableBoostingClassifier, LogisticRegression, ClassificationTree, DecisionListClassifier\n",
    "\n",
    "ebm = ExplainableBoostingClassifier(random_state=seed)\n",
    "ebm.fit(X_train, y_train)   #Works on dataframes and numpy arrays"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Global Explanations: What the model learned overall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631400566512/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631400566512/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ebm_global = ebm.explain_global(name='EBM')\n",
    "show(ebm_global)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Local Explanations: How an individual prediction was made"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631401771976/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631401771976/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ebm_local = ebm.explain_local(X_test[:5], y_test[:5], name='EBM')\n",
    "show(ebm_local)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluate EBM performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631402720616/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631402720616/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from interpret.perf import ROC\n",
    "\n",
    "ebm_perf = ROC(ebm.predict_proba).explain_perf(X_test, y_test, name='EBM')\n",
    "show(ebm_perf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Let's test out a few other Explainable Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda\\envs\\AzureML\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning:\n",
      "\n",
      "Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<interpret.glassbox.decisiontree.ClassificationTree at 0x17bd4897550>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from interpret.glassbox import LogisticRegression, ClassificationTree\n",
    "\n",
    "# We have to transform categorical variables to use Logistic Regression and Decision Tree\n",
    "X_enc = pd.get_dummies(X, prefix_sep='.')\n",
    "feature_names = list(X_enc.columns)\n",
    "X_train_enc, X_test_enc, y_train, y_test = train_test_split(X_enc, y, test_size=0.20, random_state=seed)\n",
    "\n",
    "lr = LogisticRegression(random_state=seed, feature_names=feature_names, penalty='l1')\n",
    "lr.fit(X_train_enc, y_train)\n",
    "\n",
    "tree = ClassificationTree()\n",
    "tree.fit(X_train_enc, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compare performance using the Dashboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631374188160/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631374188160/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631358642776/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631358642776/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631402720616/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631402720616/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_perf = ROC(lr.predict_proba).explain_perf(X_test_enc, y_test, name='Logistic Regression')\n",
    "tree_perf = ROC(tree.predict_proba).explain_perf(X_test_enc, y_test, name='Classification Tree')\n",
    "\n",
    "show(lr_perf)\n",
    "show(tree_perf)\n",
    "show(ebm_perf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Glassbox: All of our models have global and local explanations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631013146808/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631013146808/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631410482648/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631410482648/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631400566512/ -->\n",
       "<iframe src=\"http://127.0.0.1:7915/1631400566512/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_global = lr.explain_global(name='LR')\n",
    "tree_global = tree.explain_global(name='Tree')\n",
    "\n",
    "show(lr_global)\n",
    "show(tree_global)\n",
    "show(ebm_global)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dashboard: look at everything at once"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<!-- http://127.0.0.1:7915/1631400233480/ -->\n",
       "<a href=\"http://127.0.0.1:7915/1631400233480/\" target=\"_new\">Open in new window</a><iframe src=\"http://127.0.0.1:7915/1631400233480/\" width=100% height=800 frameBorder=\"0\"></iframe>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do everything in one shot with the InterpretML Dashboard by passing a list into show\n",
    "\n",
    "show([hist, lr_global, lr_perf, tree_global, tree_perf, ebm_global, ebm_perf], share_tables=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:AzureML]",
   "language": "python",
   "name": "conda-env-AzureML-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}