RSF-IPEC.ipynb

rsf-ipec.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "RSF-IPEC.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "mount_file_id": "1jGHxL2vL5kRTMX0kMVV9jZWDT1eXkzpH",
      "authorship_tag": "ABX9TyNw4c8ph1CcUJZyR4i4R6vR",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/alonsosilvaallende/6633f66be36d99b261d7a00f36a896f3/rsf-ipec.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LnChyaqIicM1"
      },
      "source": [
        "!cp drive/MyDrive/Time-to-Event-Analysis/util.py ."
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TZJRW0UOjBAr"
      },
      "source": [
        "!cp drive/MyDrive/Time-to-Event-Analysis/npsurvival_models.py ."
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i9ZYUFdZs8Mh",
        "outputId": "9afca690-dff1-479d-e633-01c041d57cd8"
      },
      "source": [
        "!pip install --quiet --upgrade pip\n",
        "!pip uninstall --yes --quiet osqp\n",
        "!pip install --quiet -U scikit-survival"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\n",
            "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\n",
            "  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
            "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25herror\n",
            "\u001b[33mWARNING: Discarding https://files.pythonhosted.org/packages/91/f0/047ce90bb831ab34ca287d1d23f0c61b6546cd89494566898c0e17516990/scikit-survival-0.15.0.post0.tar.gz#sha256=572c3ac6818a9d0944fc4b8176eb948051654de857e28419ecc5060bcc6fbf37 (from https://pypi.org/simple/scikit-survival/) (requires-python:>=3.7). Command errored out with exit status 1: /usr/bin/python3 /usr/local/lib/python3.7/dist-packages/pip/_vendor/pep517/in_process/_in_process.py get_requires_for_build_wheel /tmp/tmpmgszihjl Check the logs for full command output.\u001b[0m\n",
            "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "92CA1kgflbYo",
        "outputId": "57868e52-bfca-4bbd-eb62-f9f34ad08366"
      },
      "source": [
        "!pip install --quiet lifelines"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3EO7M5I2kGPq"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t1dBhvvDkQ9e"
      },
      "source": [
        "from sksurv.datasets import load_gbsg2\n",
        "\n",
        "X, y = load_gbsg2()"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wxS5EV7jkX33"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_trn, X_test, y_trn, y_test = train_test_split(X, y, random_state=42)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kLOBNaj-kfDx"
      },
      "source": [
        "scaling_cols = [c for c in X.columns if X[c].dtype.kind in ['i', 'f']]\n",
        "cat_cols = [c for c in X.columns if X[c].dtype.kind not in ['i', 'f']]"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3_K8sDKXkg8C"
      },
      "source": [
        "from sklearn.compose import ColumnTransformer\n",
        "from sklearn.preprocessing import OrdinalEncoder\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "preprocessor = ColumnTransformer(\n",
        "    [('cat-preprocessor', OrdinalEncoder(), cat_cols),\n",
        "    ('standard-scaler', StandardScaler(), scaling_cols)],\n",
        "    remainder='passthrough', sparse_threshold=0)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-fsY8FH9tcRh"
      },
      "source": [
        "# Concordance index computation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xnpbv82skp69",
        "outputId": "6a5206b9-0138-4254-a80c-00f914378e57"
      },
      "source": [
        "from sklearn.pipeline import make_pipeline\n",
        "from sksurv.ensemble import RandomSurvivalForest\n",
        "from sksurv.metrics import concordance_index_censored\n",
        "\n",
        "rsf = make_pipeline(preprocessor, RandomSurvivalForest(random_state=42))\n",
        "rsf.fit(X_trn, y_trn)\n",
        "print(f\"Concordance index: {concordance_index_censored(y_test['cens'], y_test['time'], rsf.predict(X_test))[0]:.3f}\")"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Concordance index: 0.658\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X8YMW3E9tYZB"
      },
      "source": [
        "# IPEC score computation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z8c7LOuXkxDy"
      },
      "source": [
        "from util import compute_IPEC_scores"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b-PZWT1Fqa0Z"
      },
      "source": [
        "rsf_survfunc_test = rsf.predict_survival_function(X_test, return_array=False)"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lxxyRFumlX5E",
        "outputId": "1835263b-b387-4ddc-8516-7f2937e75518"
      },
      "source": [
        "y_trn_ip = np.array([[i,j] for i, j in zip(y_trn['time'], y_trn['cens'])])\n",
        "y_test_ip = np.array([[i,j] for i, j in zip(y_test['time'], y_test['cens'])])\n",
        "times = np.concatenate((np.array([0]), rsf_survfunc_test[0].x))\n",
        "rsf_survfunc_y = \\\n",
        "[np.concatenate((np.array([1]), rsf_survfunc_test[i].y)) for i in range(len(rsf_survfunc_test))]\n",
        "tau = [times[-1]]\n",
        "\n",
        "print(f\"IPEC score: {compute_IPEC_scores(y_trn_ip, y_test_ip, times, rsf_survfunc_y, tau)[tau[0]]/tau[0]:.4f}\")"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "IPEC score: 0.3027\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}