02_SimData_True_Vs_Random.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from deap import base\n",
    "from deap import creator\n",
    "from deap import tools\n",
    "from deap import algorithms\n",
    "\n",
    "import array\n",
    "import pandas as pd\n",
    "\n",
    "import random\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from lifelines import KaplanMeierFitter\n",
    "\n",
    "from lifelines import CoxPHFitter\n",
    "\n",
    "from Data.dataset import Flchain\n",
    "from Data.dataset import SimPHDataset\n",
    "\n",
    "from scipy.stats import ttest_ind, ttest_rel\n",
    "\n",
    "from model import CPH, ExpMultiTrues\n",
    "import seaborn as sns\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset Description =======================\n",
      "Dataset Name: SimPHDataset\n",
      "Dataset Shape: (1000, 9)\n",
      "Events: 50.00 %\n",
      "NaN Values: 0.00 %\n",
      "Size and Events % in splits: (200, 50.00%), (200, 50.00%), (200, 50.00%), (200, 50.00%), (200, 50.00%), \n",
      "===========================================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "ds = SimPHDataset(percentage_cens=0.5, verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correct events_p: 0.25\n",
      "4.51 mind: 999\n",
      "Correct events_p: 0.5\n",
      "4.95 mind: 999\n",
      "Correct events_p: 0.75\n",
      "5.64 mind: 999\n"
     ]
    }
   ],
   "source": [
    "percentage_cens = 0.5\n",
    "m_exp = ExpMultiTrues(dataset_class=SimPHDataset, dataset_filepath=None, test_fract=0.3, percentage_cens=percentage_cens)\n",
    "m_exp.run(repetitions=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.save('results/true_vs_random/ctrains_true_ps.npy',np.array(m_exp.ctrains_true_ps))\n",
    "# np.save('results/true_vs_random/ctrains_some_ps.npy',np.array(m_exp.ctrains_some_ps))\n",
    "# np.save('results/true_vs_random/cvals_true_ps.npy',np.array(m_exp.cvals_true_ps))\n",
    "# np.save('results/true_vs_random/cvals_some_ps.npy',np.array(m_exp.cvals_some_ps))\n",
    "# np.save('results/true_vs_random/accuracies.npy',np.array(m_exp.accuracies))\n",
    "# np.save('results/true_vs_random/accuracies_true.npy',np.array(m_exp.accuracies_true))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "ctrains_true_ps = np.load('results/true_vs_random/ctrains_true_ps.npy')\n",
    "ctrains_some_ps = np.load('results/true_vs_random/ctrains_some_ps.npy')\n",
    "cvals_true_ps= np.load('results/true_vs_random/cvals_true_ps.npy')\n",
    "cvals_some_ps= np.load('results/true_vs_random/cvals_some_ps.npy')\n",
    "accuracies = np.load('results/true_vs_random/accuracies.npy')\n",
    "accuracies_true = np.load('results/true_vs_random/accuracies_true.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "    \n",
    "def plot_ones_acc_vs_cindex():\n",
    "    fig, ax = plt.subplots(1, 2, figsize=(15, 3))\n",
    "    ax[0].set_title('Train')\n",
    "    ax[0].scatter(accuracies_true.reshape(-1), ctrains_true_ps.reshape(-1),\n",
    "                  label='True Labels', alpha=0.2)\n",
    "    ax[0].scatter(accuracies_true.reshape(-1), ctrains_some_ps.reshape(-1),\n",
    "                  label='Random Labels', alpha=0.2)\n",
    "    # ax[0].scatter(np.array(self.accuracies_true[5:]).reshape(-1), np.array(self.ctrains_some_ps[5:]).reshape(-1), label='High True 1s', alpha=0.2)\n",
    "    # ax[0].scatter(np.array(range(1, len(self.ctrains_true_ps)+1))*0.1, self.ctrains_true_ps, marker='+', s=100, c='k')\n",
    "    ax[0].set_xlabel('E Accuracy')\n",
    "    ax[0].set_ylabel('C-Index')\n",
    "    ax[0].legend()\n",
    "\n",
    "    ax[1].set_title('Val')\n",
    "    ax[1].scatter(accuracies_true.reshape(-1), cvals_true_ps.reshape(-1), alpha=0.2,\n",
    "                  label='True Labels')\n",
    "    ax[1].scatter(accuracies_true.reshape(-1), cvals_some_ps.reshape(-1), alpha=0.2,\n",
    "                  label='Random Labels')\n",
    "    # ax[1].scatter(np.array(self.accuracies_true[5:]).reshape(-1), np.array(self.cvals_some_ps[5:]).reshape(-1), label='High True 1s', alpha=0.2)\n",
    "\n",
    "    # ax[1].scatter(np.array(range(1, len(self.cvals_true_ps)+1))*0.1, self.cvals_true_ps, marker='+', s=100, c='k')\n",
    "    ax[1].set_xlabel('E Accuracy')\n",
    "    ax[1].set_ylabel('C-Index')\n",
    "    ax[1].legend()\n",
    "    \n",
    "def plot_true_vs_random():\n",
    "    cdiff_train_ps = []\n",
    "    cdiff_val_ps = []\n",
    "    inc_prob_train_ps = []\n",
    "    inc_prob_val_ps = []\n",
    "    correctly_labeled_ps = [0.25, 0.50, 0.75]\n",
    "    for i in range(3):\n",
    "        cdiff_train = ctrains_true_ps[i] - ctrains_some_ps[i]\n",
    "        cdiff_val = cvals_true_ps[i] - cvals_some_ps[i]\n",
    "        inc_prob_train, inc_prob_val = (cdiff_train > 0).mean(), (cdiff_val > 0).mean()\n",
    "        cdiff_train_ps.append(cdiff_train)\n",
    "        cdiff_val_ps.append(cdiff_val)\n",
    "        inc_prob_train_ps.append(inc_prob_train)\n",
    "        inc_prob_val_ps.append(inc_prob_val)\n",
    "\n",
    "        fig, ax = plt.subplots(1, 3, figsize=(15, 3))\n",
    "        # 'Correctly Labeled Events: {:.1f}'.format(((i+1)*0.1))\n",
    "        fig.suptitle(f'Censoring:{percentage_cens:.1f}, Correctly labeled event {correctly_labeled_ps[i]:.1f}')\n",
    "        ax[0].set_title('Train C-Index')\n",
    "        ax[0].plot(ctrains_true_ps[i], label='True Events')\n",
    "        ax[0].plot(ctrains_some_ps[i], label='Random Events')\n",
    "        ax[0].set_xlabel('Iterations')\n",
    "        ax[0].set_ylabel('CI')\n",
    "        ax[0].legend()\n",
    "\n",
    "        ax[1].set_title('Val C-Index')\n",
    "        ax[1].plot(cvals_true_ps[i], label='True Events')\n",
    "        ax[1].plot(cvals_some_ps[i], label='Random Events')\n",
    "        ax[1].set_xlabel('Iterations')\n",
    "        ax[1].set_ylabel('CI')\n",
    "        ax[1].legend()\n",
    "\n",
    "        ax[2].set_title('P(true>random), Train: {:.2f}, Val: {:.2f}'.format(inc_prob_train, inc_prob_val))\n",
    "        ax[2].plot(cdiff_train, label='Train C-Index Diff', c='k')\n",
    "        ax[2].plot(cdiff_val, label='Val C-Index Diff', c='gray')\n",
    "        plt.axhline(0, c='C2', ls=':')\n",
    "        ax[2].set_xlabel('Iterations')\n",
    "        ax[2].set_ylabel('CI diff')\n",
    "        ax[2].legend()\n",
    "        fig.tight_layout()\n",
    "\n",
    "        \n",
    "def plot_true_vs_random_box():\n",
    "    correctly_labeled_ps = [0.25, 0.50, 0.75]\n",
    "    for i in range(3):\n",
    "        #fig, ax = plt.subplots(1, 2, figsize=(6, 2))\n",
    "        \n",
    "        cdiff_train = ctrains_true_ps[i] - ctrains_some_ps[i]\n",
    "        cdiff_val = cvals_true_ps[i] - cvals_some_ps[i]\n",
    "        inc_prob_train, inc_prob_val = (cdiff_train > 0).mean(), (cdiff_val > 0).mean()\n",
    "        \n",
    "        plt.figure(figsize=(3,2))\n",
    "        #plt.boxplot([ctrains_true_ps[i], ctrains_some_ps[i]]);\n",
    "        sns.boxplot(data=[ctrains_true_ps[i], ctrains_some_ps[i]]);\n",
    "        plt.xticks([0, 1], ['True', 'Random'])\n",
    "        plt.ylabel('C-Index')\n",
    "        plt.title('P(true>random): {:.2f}'.format(inc_prob_train))\n",
    "        plt.ylim(0.7, 1)\n",
    "        plt.tight_layout()\n",
    "        plt.savefig(f'results/true_vs_random/figs/correctly_labeled_p_{correctly_labeled_ps[i]}_train.pdf')\n",
    "        \n",
    "        plt.figure(figsize=(3,2))\n",
    "        #plt.boxplot([cvals_true_ps[i], cvals_some_ps[i]]);\n",
    "        sns.boxplot(data=[cvals_true_ps[i], cvals_some_ps[i]]);\n",
    "        plt.xticks([0, 1], ['True', 'Random'])\n",
    "        plt.ylabel('C-Index')\n",
    "        plt.title('P(true>random): {:.2f}'.format(inc_prob_val))\n",
    "        plt.ylim(0.7, 1)\n",
    "        plt.tight_layout()\n",
    "        plt.savefig(f'results/true_vs_random/figs/correctly_labeled_p_{correctly_labeled_ps[i]}_test.pdf')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_true_vs_random_box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_true_vs_random_box():\n",
    "    correctly_labeled_ps = [0.25, 0.50, 0.75]\n",
    "    for i in range(3):\n",
    "        #fig, ax = plt.subplots(1, 2, figsize=(6, 2))\n",
    "        \n",
    "        cdiff_train = ctrains_true_ps[i] - ctrains_some_ps[i]\n",
    "        cdiff_val = cvals_true_ps[i] - cvals_some_ps[i]\n",
    "        inc_prob_train, inc_prob_val = (cdiff_train > 0).mean(), (cdiff_val > 0).mean()\n",
    "        \n",
    "        plt.figure(figsize=(3,2))\n",
    "        #plt.boxplot([ctrains_true_ps[i], ctrains_some_ps[i]]);\n",
    "        sns.boxplot(data=[ctrains_true_ps[i], ctrains_some_ps[i]]);\n",
    "        plt.xticks([0, 1], ['True', 'Random'])\n",
    "        plt.ylabel('C-Index')\n",
    "        plt.title('P(true>random): {:.2f}'.format(inc_prob_train))\n",
    "        plt.ylim(0.7, 1)\n",
    "        plt.tight_layout()\n",
    "        plt.savefig(f'results/true_vs_random/figs/correctly_labeled_p_{correctly_labeled_ps[i]}_train.pdf')\n",
    "        \n",
    "        plt.figure(figsize=(3,2))\n",
    "        #plt.boxplot([cvals_true_ps[i], cvals_some_ps[i]]);\n",
    "        sns.boxplot(data=[cvals_true_ps[i], cvals_some_ps[i]]);\n",
    "        plt.xticks([0, 1], ['True', 'Random'])\n",
    "        plt.ylabel('C-Index')\n",
    "        plt.title('P(true>random): {:.2f}'.format(inc_prob_val))\n",
    "        plt.ylim(0.7, 1)\n",
    "        plt.tight_layout()\n",
    "        plt.savefig(f'results/true_vs_random/figs/correctly_labeled_p_{correctly_labeled_ps[i]}_test.pdf')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train_true = pd.DataFrame(ctrains_true_ps.transpose(), columns=['25%', '50%', '75%'])\n",
    "df_train_true['Label'] = 'True'\n",
    "df_train_random = pd.DataFrame(ctrains_some_ps.transpose(), columns=['25%', '50%', '75%'])\n",
    "df_train_random['Label'] = 'Random'\n",
    "\n",
    "df_train = pd.concat([df_train_true, df_train_random])\n",
    "\n",
    "\n",
    "df_test_true = pd.DataFrame(cvals_true_ps.transpose(), columns=['25%', '50%', '75%'])\n",
    "df_test_true['Label'] = 'True'\n",
    "df_test_random = pd.DataFrame(cvals_some_ps.transpose(), columns=['25%', '50%', '75%'])\n",
    "df_test_random['Label'] = 'Random'\n",
    "\n",
    "df_test = pd.concat([df_test_true, df_test_random])\n",
    "\n",
    "df_train_m = pd.melt(df_train, id_vars=['Label'], value_vars=['25%', '50%', '75%'])\n",
    "df_test_m = pd.melt(df_test, id_vars=['Label'], value_vars=['25%', '50%', '75%'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(4,3))\n",
    "sns.boxplot(data=df_train_m, x='variable', y='value', hue='Label');\n",
    "plt.xlabel('Correct Events %')\n",
    "plt.ylabel('C-Index')\n",
    "plt.tight_layout()\n",
    "plt.savefig(f'results/true_vs_random/figs/true_vs_random_train.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(4,3))\n",
    "sns.boxplot(data=df_test_m, x='variable', y='value', hue='Label');\n",
    "plt.xlabel('Correct Events %')\n",
    "plt.ylabel('C-Index')\n",
    "plt.tight_layout()\n",
    "plt.savefig(f'results/true_vs_random/figs/true_vs_random_test.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}