examples/Model_EfnBoost.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## EfnBoost for Survival Analysis\n",
    "\n",
    "### Introduction\n",
    "EfnBoost is a method of survival analysis that predicts hazard ratio of coxph using `xgboost` by optimizing the loss function of *Efron Approximation*.\n",
    "\n",
    "Here we show you how to use EfnBoost model to predict hazard ratio from scratch with the help of `libsurv`.\n",
    "\n",
    "### Preparation\n",
    "\n",
    "For all things going well, you would be better to get acquainted with **Survival Analysis**. Otherwise, I suggest you read the [reference](https://lifelines.readthedocs.io/en/latest/Survival%20Analysis%20intro.html).\n",
    "\n",
    "And it may help a lot if you know a little of the Efron Approximation.\n",
    "\n",
    "### Get it started\n",
    "\n",
    "#### Load datasets\n",
    "\n",
    "We provide `WHAS` and `METABRIC` datasets for further studies. It can be found in module `libsurv.datasets`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from libsurv.datasets import load_whas_train\n",
    "from libsurv.datasets import load_whas_test\n",
    "\n",
    "data_train = load_whas_train()\n",
    "data_test = load_whas_test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['age', 'gender', 'bmi', 'chf', 'miord', 'e', 't'], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(data_train.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Statistics of datasets\n",
    "Print statistics of survival data. The function `survival_stats` in `libsurv.datasets` provides the tool."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Survival Data Statistics ---------------\n",
      "# Rows: 1310\n",
      "# Columns: 5 + e + t\n",
      "# Events Ratio: 42.14%\n",
      "# Min Time: 1\n",
      "# Max Time: 67\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from libsurv.datasets import survival_stats\n",
    "\n",
    "survival_stats(data_train, t_col=\"t\", e_col=\"e\", plot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Survival Data Statistics ---------------\n",
      "# Rows: 328\n",
      "# Columns: 5 + e + t\n",
      "# Events Ratio: 42.07%\n",
      "# Min Time: 1\n",
      "# Max Time: 67\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "survival_stats(data_test, t_col=\"t\", e_col=\"e\", plot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Survival data transfrom\n",
    "\n",
    "The column of time and status in survival data will be packed into an new column as label in the transformed data.\n",
    "\n",
    "The transformed survival data contains an new label, where negtive values are considered as right censored, and positive values are considered as event occurrence.\n",
    "\n",
    "**NOTE**: In `libsurv`, survival data must be transformed via `libsurv.datasets.survival_df`. But for `xgboost` model, the type of data must be `DMatrix`, so we **have two ways to achive this as listed below**.\n",
    "\n",
    "```python\n",
    "##############################################################################\n",
    "# method 1:\n",
    "from libsurv.datasets import survival_dmat\n",
    "\n",
    "surv_train = survival_dmat(data_train, t_col=\"t\", e_col=\"e\", label_col=\"Y\")\n",
    "surv_test = survival_dmat(data_test, t_col=\"t\", e_col=\"e\", label_col=\"Y\")\n",
    "\n",
    "##############################################################################\n",
    "\n",
    "##############################################################################\n",
    "# method 2\n",
    "from libsurv.datasets import survival_df\n",
    "from xgboost import DMatrix\n",
    "\n",
    "\n",
    "data_train = survival_df(data_train, t_col=\"t\", e_col=\"e\", label_col=\"Y\")\n",
    "data_test = survival_df(data_test, t_col=\"t\", e_col=\"e\", label_col=\"Y\")\n",
    "\n",
    "x_cols = list(data_train.columns)[:-1]\n",
    "\n",
    "surv_train = DMatrix(data_train[x_cols], label=data_train['Y'].values)\n",
    "surv_test = DMatrix(data_test[x_cols], label=data_test['Y'].values)\n",
    "################################################################################\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from libsurv.datasets import survival_dmat\n",
    "\n",
    "# via survival_dmat function\n",
    "surv_train = survival_dmat(data_train, t_col=\"t\", e_col=\"e\", label_col=\"Y\")\n",
    "surv_test = survival_dmat(data_test, t_col=\"t\", e_col=\"e\", label_col=\"Y\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model initialization\n",
    "\n",
    "Here we start to initialize a EfnBoost model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from libsurv import EfnBoost\n",
    "\n",
    "params = {\n",
    "    'eta':0.3,\n",
    "    'max_depth':6, \n",
    "    'min_child_weight': 8, \n",
    "    'subsample': 0.9,\n",
    "    'colsample_bytree': 0.5,\n",
    "    'gamma': 0.20,\n",
    "    'lambda': 0,\n",
    "}\n",
    "\n",
    "model = EfnBoost(params)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model training\n",
    "\n",
    "Training model and plotting the learning curve."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\ttrain-rmse:41.3548\ttest-rmse:41.5541\ttrain-efron_loss:6.40707\ttest-efron_loss:5.04637\n",
      "[10]\ttrain-rmse:40.8382\ttest-rmse:41.0603\ttrain-efron_loss:5.81787\ttest-efron_loss:4.61236\n",
      "[20]\ttrain-rmse:40.6026\ttest-rmse:40.8388\ttrain-efron_loss:5.63601\ttest-efron_loss:4.52832\n",
      "[30]\ttrain-rmse:40.4514\ttest-rmse:40.6784\ttrain-efron_loss:5.53066\ttest-efron_loss:4.44715\n",
      "[40]\ttrain-rmse:40.3762\ttest-rmse:40.5953\ttrain-efron_loss:5.45829\ttest-efron_loss:4.43786\n",
      "[50]\ttrain-rmse:40.2692\ttest-rmse:40.5002\ttrain-efron_loss:5.39953\ttest-efron_loss:4.37979\n",
      "[60]\ttrain-rmse:40.1823\ttest-rmse:40.4041\ttrain-efron_loss:5.33884\ttest-efron_loss:4.3381\n",
      "[70]\ttrain-rmse:40.1277\ttest-rmse:40.3542\ttrain-efron_loss:5.28602\ttest-efron_loss:4.3264\n",
      "[80]\ttrain-rmse:40.0519\ttest-rmse:40.271\ttrain-efron_loss:5.22615\ttest-efron_loss:4.30638\n",
      "[90]\ttrain-rmse:40.0328\ttest-rmse:40.2443\ttrain-efron_loss:5.20025\ttest-efron_loss:4.2888\n",
      "[99]\ttrain-rmse:40.0182\ttest-rmse:40.2431\ttrain-efron_loss:5.17733\ttest-efron_loss:4.28427\n"
     ]
    },
    {
     "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": [
    "eval_result = model.train(\n",
    "    surv_train,\n",
    "    num_rounds=100,\n",
    "    skip_rounds=10,\n",
    "    evals=[(surv_train, 'train'), (surv_test, 'test')],\n",
    "    plot=True\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Prediction and Evaluation\n",
    "\n",
    "Now we use the trained model to make prediction or evaluate model performance.\n",
    "\n",
    "Give the prediction of EfnBoost model with:\n",
    "- `model.predict`: Hazard Ratio of coxph\n",
    "- `model.predict_survival_function`: survival function transformed from hazard ratio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction of hazard ratio: [0.0694712  7.574963   0.38022602 0.2501495  7.308943  ]\n"
     ]
    },
    {
     "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": [
    "# select the first 5 samples\n",
    "surv_data_now = surv_test.slice([i for i in range(5)])\n",
    "\n",
    "# Predict Hazard Ratio\n",
    "print(\"Prediction of hazard ratio:\", model.predict(surv_data_now, output_margin=False))\n",
    "\n",
    "# Predict Survival Function\n",
    "result_survf = model.predict_survival_function(surv_data_now, plot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>1.0</th>\n",
       "      <th>2.0</th>\n",
       "      <th>3.0</th>\n",
       "      <th>4.0</th>\n",
       "      <th>5.0</th>\n",
       "      <th>6.0</th>\n",
       "      <th>7.0</th>\n",
       "      <th>8.0</th>\n",
       "      <th>9.0</th>\n",
       "      <th>10.0</th>\n",
       "      <th>...</th>\n",
       "      <th>58.0</th>\n",
       "      <th>59.0</th>\n",
       "      <th>60.0</th>\n",
       "      <th>61.0</th>\n",
       "      <th>62.0</th>\n",
       "      <th>63.0</th>\n",
       "      <th>64.0</th>\n",
       "      <th>65.0</th>\n",
       "      <th>66.0</th>\n",
       "      <th>67.0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.999449</td>\n",
       "      <td>0.999152</td>\n",
       "      <td>0.999062</td>\n",
       "      <td>0.998829</td>\n",
       "      <td>0.998717</td>\n",
       "      <td>0.998546</td>\n",
       "      <td>0.998349</td>\n",
       "      <td>0.998288</td>\n",
       "      <td>0.998180</td>\n",
       "      <td>0.998002</td>\n",
       "      <td>...</td>\n",
       "      <td>0.959234</td>\n",
       "      <td>0.957893</td>\n",
       "      <td>0.951494</td>\n",
       "      <td>0.948184</td>\n",
       "      <td>0.948184</td>\n",
       "      <td>0.948184</td>\n",
       "      <td>0.938714</td>\n",
       "      <td>0.929152</td>\n",
       "      <td>8.808165e-01</td>\n",
       "      <td>8.808165e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.941627</td>\n",
       "      <td>0.911646</td>\n",
       "      <td>0.902757</td>\n",
       "      <td>0.880069</td>\n",
       "      <td>0.869350</td>\n",
       "      <td>0.853320</td>\n",
       "      <td>0.835093</td>\n",
       "      <td>0.829564</td>\n",
       "      <td>0.819855</td>\n",
       "      <td>0.804103</td>\n",
       "      <td>...</td>\n",
       "      <td>0.010693</td>\n",
       "      <td>0.009180</td>\n",
       "      <td>0.004420</td>\n",
       "      <td>0.003023</td>\n",
       "      <td>0.003023</td>\n",
       "      <td>0.003023</td>\n",
       "      <td>0.001012</td>\n",
       "      <td>0.000331</td>\n",
       "      <td>9.782502e-07</td>\n",
       "      <td>9.782502e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.996986</td>\n",
       "      <td>0.995368</td>\n",
       "      <td>0.994878</td>\n",
       "      <td>0.993608</td>\n",
       "      <td>0.992997</td>\n",
       "      <td>0.992070</td>\n",
       "      <td>0.990995</td>\n",
       "      <td>0.990665</td>\n",
       "      <td>0.990079</td>\n",
       "      <td>0.989116</td>\n",
       "      <td>...</td>\n",
       "      <td>0.796289</td>\n",
       "      <td>0.790215</td>\n",
       "      <td>0.761752</td>\n",
       "      <td>0.747362</td>\n",
       "      <td>0.747362</td>\n",
       "      <td>0.747362</td>\n",
       "      <td>0.707408</td>\n",
       "      <td>0.668859</td>\n",
       "      <td>4.992868e-01</td>\n",
       "      <td>4.992868e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.998016</td>\n",
       "      <td>0.996950</td>\n",
       "      <td>0.996627</td>\n",
       "      <td>0.995790</td>\n",
       "      <td>0.995387</td>\n",
       "      <td>0.994776</td>\n",
       "      <td>0.994066</td>\n",
       "      <td>0.993848</td>\n",
       "      <td>0.993462</td>\n",
       "      <td>0.992826</td>\n",
       "      <td>...</td>\n",
       "      <td>0.860825</td>\n",
       "      <td>0.856499</td>\n",
       "      <td>0.836076</td>\n",
       "      <td>0.825651</td>\n",
       "      <td>0.825651</td>\n",
       "      <td>0.825651</td>\n",
       "      <td>0.796340</td>\n",
       "      <td>0.767518</td>\n",
       "      <td>6.332066e-01</td>\n",
       "      <td>6.332066e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.943618</td>\n",
       "      <td>0.914612</td>\n",
       "      <td>0.906006</td>\n",
       "      <td>0.884026</td>\n",
       "      <td>0.873635</td>\n",
       "      <td>0.858087</td>\n",
       "      <td>0.840395</td>\n",
       "      <td>0.835025</td>\n",
       "      <td>0.825594</td>\n",
       "      <td>0.810283</td>\n",
       "      <td>...</td>\n",
       "      <td>0.012541</td>\n",
       "      <td>0.010824</td>\n",
       "      <td>0.005347</td>\n",
       "      <td>0.003706</td>\n",
       "      <td>0.003706</td>\n",
       "      <td>0.003706</td>\n",
       "      <td>0.001289</td>\n",
       "      <td>0.000439</td>\n",
       "      <td>1.590360e-06</td>\n",
       "      <td>1.590360e-06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 61 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       1.0       2.0       3.0       4.0       5.0       6.0       7.0   \\\n",
       "0  0.999449  0.999152  0.999062  0.998829  0.998717  0.998546  0.998349   \n",
       "1  0.941627  0.911646  0.902757  0.880069  0.869350  0.853320  0.835093   \n",
       "2  0.996986  0.995368  0.994878  0.993608  0.992997  0.992070  0.990995   \n",
       "3  0.998016  0.996950  0.996627  0.995790  0.995387  0.994776  0.994066   \n",
       "4  0.943618  0.914612  0.906006  0.884026  0.873635  0.858087  0.840395   \n",
       "\n",
       "       8.0       9.0       10.0  ...      58.0      59.0      60.0      61.0  \\\n",
       "0  0.998288  0.998180  0.998002  ...  0.959234  0.957893  0.951494  0.948184   \n",
       "1  0.829564  0.819855  0.804103  ...  0.010693  0.009180  0.004420  0.003023   \n",
       "2  0.990665  0.990079  0.989116  ...  0.796289  0.790215  0.761752  0.747362   \n",
       "3  0.993848  0.993462  0.992826  ...  0.860825  0.856499  0.836076  0.825651   \n",
       "4  0.835025  0.825594  0.810283  ...  0.012541  0.010824  0.005347  0.003706   \n",
       "\n",
       "       62.0      63.0      64.0      65.0          66.0          67.0  \n",
       "0  0.948184  0.948184  0.938714  0.929152  8.808165e-01  8.808165e-01  \n",
       "1  0.003023  0.003023  0.001012  0.000331  9.782502e-07  9.782502e-07  \n",
       "2  0.747362  0.747362  0.707408  0.668859  4.992868e-01  4.992868e-01  \n",
       "3  0.825651  0.825651  0.796340  0.767518  6.332066e-01  6.332066e-01  \n",
       "4  0.003706  0.003706  0.001289  0.000439  1.590360e-06  1.590360e-06  \n",
       "\n",
       "[5 rows x 61 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_survf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "td-CI on training set: 0.9256697500240054\n",
      "td-CI on test set: 0.856783834267278\n"
     ]
    }
   ],
   "source": [
    "# evaluate model performance \n",
    "print(\"td-CI on training set:\", model.evals(surv_train))\n",
    "print(\"td-CI on test set:\", model.evals(surv_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Factor improtance\n",
    "\n",
    "Get the factor importance score evaluated by the model.\n",
    "\n",
    "It's suggested that you **repeat obtaining the factor score for multiply times**, such as 20, by specifing a different random seed in `model_params`. \n",
    "\n",
    "Parameter `importance_type`: str. The metrics of importance evaluation. see more in [https://xgboost.readthedocs.io/en/latest/python/python_api.html#xgboost.Booster.get_score].\n",
    "\n",
    "The function returns a dict."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'bmi': 0.5448776065276518, 'miord': 0.04533091568449683, 'gender': 0.04170444242973708, 'age': 0.3327289211242067, 'chf': 0.03535811423390753}\n"
     ]
    }
   ],
   "source": [
    "fscore = model.get_factor_score(importance_type='weight')\n",
    "\n",
    "# normalize the scores\n",
    "sum_score = sum(fscore.values())\n",
    "fscore = {k: v / sum_score for k, v in fscore.items()}\n",
    "print(fscore)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### save model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.save_model(\"EfnBoost.model\")"
   ]
  }
 ],
 "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": "3.6.2"
  },
  "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": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}