02-1_plot_run_info_stat_iccm2019.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from run_info_utils import *\n",
    "from utils import create_dir\n",
    "import config\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from os.path import join"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "experiment_name = 'iccm2019'\n",
    "plot_dir = 'plot_figures'\n",
    "fig_ext = 'pdf'\n",
    "all_correct_case = 'init_all_correct'  # 'init_complete_all_correct', 'init_all_correct'\n",
    "font_size = {'xlabel':20, 'ylabel':20, 'xtick':18, 'ytick':18}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_fig(fig_name, experiment_name):\n",
    "    plot_fig_path = '{}/{}/{}.{}'.format(plot_dir, experiment_name, fig_name, fig_ext)\n",
    "    create_dir('{}/{}'.format(plot_dir, experiment_name))\n",
    "    plt.savefig(plot_fig_path, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Mean Answer Step By Carries"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.1. Mean Answer Step By Carries: Addition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean_answer_step_by_carries_add_cp7_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp7-h24.csv'))\n",
    "df_mean_answer_step_by_carries_add_cp7_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp7-h48.csv'))\n",
    "df_mean_answer_step_by_carries_add_cp7_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp7-h72.csv'))\n",
    "\n",
    "df_mean_answer_step_by_carries_add_cp8_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp8-h24.csv'))\n",
    "df_mean_answer_step_by_carries_add_cp8_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp8-h48.csv'))\n",
    "df_mean_answer_step_by_carries_add_cp8_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp8-h72.csv'))\n",
    "\n",
    "df_mean_answer_step_by_carries_add_cp9_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp9-h24.csv'))\n",
    "df_mean_answer_step_by_carries_add_cp9_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp9-h48.csv'))\n",
    "df_mean_answer_step_by_carries_add_cp9_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-add-cp9-h72.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 493.714x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df_mean_answer_step_by_carries_add_cp7_h24['carries'].unique()\n",
    "\n",
    "n_carry_types = 5\n",
    "plt.figure(figsize=(2 * 6 / 7 * (n_carry_types-1), 6))\n",
    "plt.xlabel('#Carries', fontsize=font_size['xlabel'])\n",
    "plt.ylabel('Mean answer step', fontsize=font_size['ylabel'])\n",
    "plt.tick_params(axis='x', labelsize=font_size['xtick'])\n",
    "plt.tick_params(axis='y', labelsize=font_size['ytick'])\n",
    "plt.xticks(x)\n",
    "plt.ylim(bottom=0, top=6)  \n",
    "plt.grid(axis='y')\n",
    "\n",
    "markersize = 12\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp7_h24.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp7_h24.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='b', capsize=3)\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='*', markersize=markersize, label=r'$\\mathtt{\\theta}$7d24')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp7_h48.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp7_h48.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='b', capsize=3)\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$7d48')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp7_h72.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp7_h72.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='b', capsize=3)\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='x', markersize=markersize, label=r'$\\mathtt{\\theta}$7d72')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp8_h24.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp8_h24.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='g', capsize=3)\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='*', markersize=markersize, label=r'$\\mathtt{\\theta}$8d24')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp8_h48.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp8_h48.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='g', capsize=3)\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$8d48')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp8_h72.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp8_h72.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='g', capsize=3)\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='x', markersize=markersize, label=r'$\\mathtt{\\theta}$8d72')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp9_h24.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp9_h24.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='#cc0000', capsize=18)\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='*', markersize=markersize, label=r'$\\mathtt{\\theta}$9d24')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp9_h48.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp9_h48.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='none', ecolor='#cc0000', capsize=27)\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$9d48')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_add_cp9_h72.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_add_cp9_h72.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, np.zeros_like(e), e, fmt='none', ecolor='#cc0000', capsize=36)\n",
    "plt.errorbar(x, y, e, fmt='none', ecolor='orange', capsize=6)\n",
    "\n",
    "i=0; plt.text(x[i]-0.1, y[i]+e[i]+0.1, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='left', rotation=0)\n",
    "i=1; plt.text(x[i], y[i]+e[i]+0.2, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='center', rotation=0)\n",
    "i=2; plt.text(x[i], y[i]+e[i]+0.2, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='center', rotation=0)\n",
    "i=3; plt.text(x[i], y[i]+e[i]+0.1, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='center', rotation=0)\n",
    "i=4; plt.text(x[i]+0.1, y[i]+e[i]+0.3, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='right', rotation=0)\n",
    "\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='x', markersize=markersize, label=r'$\\mathtt{\\theta}$9d72')\n",
    "\n",
    "plt.legend(ncol=3, loc='upper left', fontsize=15)\n",
    "\n",
    "save_fig('mean_answer_step_by_carries-add', 'iccm2019')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.2. Mean Answer Step By Carries: Subtract"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean_answer_step_by_carries_sub_cp7_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp7-h24.csv'))\n",
    "df_mean_answer_step_by_carries_sub_cp7_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp7-h48.csv'))\n",
    "df_mean_answer_step_by_carries_sub_cp7_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp7-h72.csv'))\n",
    "\n",
    "df_mean_answer_step_by_carries_sub_cp8_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp8-h24.csv'))\n",
    "df_mean_answer_step_by_carries_sub_cp8_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp8-h48.csv'))\n",
    "df_mean_answer_step_by_carries_sub_cp8_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp8-h72.csv'))\n",
    "\n",
    "df_mean_answer_step_by_carries_sub_cp9_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp9-h24.csv'))\n",
    "df_mean_answer_step_by_carries_sub_cp9_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp9-h48.csv'))\n",
    "df_mean_answer_step_by_carries_sub_cp9_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_carries-sub-cp9-h72.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 370.286x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df_mean_answer_step_by_carries_sub_cp7_h24['carries'].unique()\n",
    "\n",
    "n_carry_types = 4\n",
    "plt.figure(figsize=(2 * 6 / 7 * (n_carry_types-1), 6))\n",
    "plt.xlabel('#Carries', fontsize=font_size['xlabel'])\n",
    "plt.ylabel('Mean answer step', fontsize=font_size['ylabel'])\n",
    "plt.tick_params(axis='x', labelsize=font_size['xtick'])\n",
    "plt.tick_params(axis='y', labelsize=font_size['ytick'])\n",
    "plt.xticks(x)\n",
    "plt.ylim(bottom=0, top=6)  \n",
    "plt.grid(axis='y')\n",
    "\n",
    "markersize = 12\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp7_h24.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp7_h24.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7_h24')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='*', markersize=markersize, label=r'$\\mathtt{\\theta}$7d24')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp7_h48.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp7_h48.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7_h48')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$7d48')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp7_h72.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp7_h72.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7_h72')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='x', markersize=markersize, label=r'$\\mathtt{\\theta}$7d72')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp8_h24.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp8_h24.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='b:o', ecolor='orange', capsize=3, label='cp8_h24')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='*', markersize=markersize, label=r'$\\mathtt{\\theta}$8d24')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp8_h48.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp8_h48.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='b:o', ecolor='orange', capsize=3, label='cp8_h48')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$8d48')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp8_h72.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp8_h72.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='b:o', ecolor='orange', capsize=3, label='cp8_h72')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='x', markersize=markersize, label=r'$\\mathtt{\\theta}$8d72')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp9_h24.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp9_h24.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='g:o', ecolor='orange', capsize=3, label='cp9_h24')\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='*', markersize=markersize, label=r'$\\mathtt{\\theta}$9d24')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp9_h48.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp9_h48.groupby('carries').std()['mean_answer_steps'].values\n",
    "#plt.errorbar(x, y, e, fmt='g:o', ecolor='orange', capsize=3, label='cp9_h48')\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$9d48')\n",
    "\n",
    "y = df_mean_answer_step_by_carries_sub_cp9_h72.groupby('carries').mean()['mean_answer_steps'].values\n",
    "e = df_mean_answer_step_by_carries_sub_cp9_h72.groupby('carries').std()['mean_answer_steps'].values\n",
    "plt.errorbar(x, y, e, fmt='none', ecolor='orange', capsize=6)\n",
    "\n",
    "i=0; plt.text(x[i], y[i]+e[i]+0.1, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='left', rotation=30, rotation_mode='anchor')\n",
    "i=1; plt.text(x[i], y[i]+e[i]+0.4, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='center', rotation=0)\n",
    "i=2; plt.text(x[i]-0.1, y[i]+e[i], 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='right', verticalalignment='center', rotation=0)\n",
    "i=3; plt.text(x[i]-0.1, y[i]-e[i]+0.1, 'SD={0:.2f}'.format(round(e[i],2)), fontsize=font_size['xtick'], color='orange', horizontalalignment='right', verticalalignment='center', rotation=33, rotation_mode='anchor')\n",
    "\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='x', markersize=markersize, label=r'$\\mathtt{\\theta}$9d72')\n",
    "\n",
    "plt.legend(ncol=3, loc='upper left', fontsize=15, handletextpad=0.1, columnspacing=0.1)\n",
    "\n",
    "save_fig('mean_answer_step_by_carries-sub', 'iccm2019')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Mean Answer Step By Confidence Probability"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.1. Mean Answer Step By Confidence Probability: Addition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean_answer_step_by_cp_add_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_cp-add-h24.csv'))\n",
    "df_mean_answer_step_by_cp_add_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_cp-add-h48.csv'))\n",
    "df_mean_answer_step_by_cp_add_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_cp-add-h72.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df_mean_answer_step_by_cp_add_h24['confidence_prob'].unique()\n",
    "x.sort()\n",
    "\n",
    "ylimit = 3\n",
    "\n",
    "plt.figure(figsize=(6, ylimit))\n",
    "plt.xlabel(r'Confidence threshold ($\\theta_{c}$)', fontsize=font_size['xlabel'])\n",
    "plt.ylabel('Mean answer step', fontsize=font_size['ylabel'])\n",
    "plt.tick_params(axis='x', labelsize=font_size['xtick'])\n",
    "plt.tick_params(axis='y', labelsize=font_size['ytick'])\n",
    "plt.xticks(x)\n",
    "plt.yticks(np.arange(0,ylimit+0.1,1))\n",
    "plt.ylim(bottom=0, top=ylimit)  \n",
    "plt.grid(axis='y')\n",
    "\n",
    "markersize = 12\n",
    "\n",
    "y = df_mean_answer_step_by_cp_add_h24.groupby('confidence_prob').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_cp_add_h24.groupby('confidence_prob').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7_h24')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='.', markersize=markersize, label=r'd24')\n",
    "\n",
    "y = df_mean_answer_step_by_cp_add_h48.groupby('confidence_prob').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_cp_add_h48.groupby('confidence_prob').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7_h24')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='.', markersize=markersize, label=r'd48')\n",
    "\n",
    "y = df_mean_answer_step_by_cp_add_h72.groupby('confidence_prob').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_cp_add_h72.groupby('confidence_prob').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7_h24')\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='.', markersize=markersize, label=r'd72')\n",
    "\n",
    "plt.legend(ncol=3, loc='upper left', fontsize=15)\n",
    "\n",
    "save_fig('mean_answer_step_by_cp-add', 'iccm2019')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.2. Mean Answer Step By Confidence Probability: Subtraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean_answer_step_by_cp_sub_h24 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_cp-sub-h24.csv'))\n",
    "df_mean_answer_step_by_cp_sub_h48 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_cp-sub-h48.csv'))\n",
    "df_mean_answer_step_by_cp_sub_h72 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_cp-sub-h72.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df_mean_answer_step_by_cp_sub_h24['confidence_prob'].unique()\n",
    "x.sort()\n",
    "\n",
    "ylimit = 3\n",
    "\n",
    "plt.figure(figsize=(6, ylimit))\n",
    "plt.xlabel(r'Confidence threshold ($\\theta_{c}$)', fontsize=font_size['xlabel'])\n",
    "plt.ylabel('Mean answer step', fontsize=font_size['ylabel'])\n",
    "plt.tick_params(axis='x', labelsize=font_size['xtick'])\n",
    "plt.tick_params(axis='y', labelsize=font_size['ytick'])\n",
    "plt.xticks(x)\n",
    "plt.yticks(np.arange(0,ylimit+0.1,1))\n",
    "plt.ylim(bottom=0, top=ylimit)  \n",
    "plt.grid(axis='y')\n",
    "\n",
    "markersize = 12\n",
    "\n",
    "y = df_mean_answer_step_by_cp_sub_h24.groupby('confidence_prob').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_cp_sub_h24.groupby('confidence_prob').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='h24')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='.', markersize=markersize, label=r'd24')\n",
    "\n",
    "y = df_mean_answer_step_by_cp_sub_h48.groupby('confidence_prob').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_cp_sub_h48.groupby('confidence_prob').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='h48')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='.', markersize=markersize, label=r'd48')\n",
    "\n",
    "y = df_mean_answer_step_by_cp_sub_h72.groupby('confidence_prob').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_cp_sub_h72.groupby('confidence_prob').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='h72')\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='.', markersize=markersize, label=r'd72')\n",
    "\n",
    "plt.legend(ncol=3, loc='upper left', fontsize=15)\n",
    "\n",
    "save_fig('mean_answer_step_by_cp-sub', 'iccm2019')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Mean Answer Step By Hidden Dimension"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.1. Mean Answer Step By Hidden Dimension: Addition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean_answer_step_by_hd_add_cp7 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_hd-add-cp7.csv'))\n",
    "df_mean_answer_step_by_hd_add_cp8 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_hd-add-cp8.csv'))\n",
    "df_mean_answer_step_by_hd_add_cp9 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_hd-add-cp9.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df_mean_answer_step_by_hd_add_cp7['hidden_dimension'].unique()\n",
    "x.sort()\n",
    "\n",
    "ylimit = 3\n",
    "\n",
    "plt.figure(figsize=(6, ylimit))\n",
    "plt.xlabel(r'Hidden dimension ($d_c$)', fontsize=font_size['xlabel'])\n",
    "plt.ylabel('Mean answer step', fontsize=font_size['ylabel'])\n",
    "plt.tick_params(axis='x', labelsize=font_size['xtick'])\n",
    "plt.tick_params(axis='y', labelsize=font_size['ytick'])\n",
    "plt.xticks(x)\n",
    "plt.yticks(np.arange(0,ylimit+0.1,1))\n",
    "plt.ylim(bottom=0, top=ylimit)  \n",
    "plt.grid(axis='y')\n",
    "\n",
    "markersize = 12\n",
    "\n",
    "y = df_mean_answer_step_by_hd_add_cp7.groupby('hidden_dimension').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_hd_add_cp7.groupby('hidden_dimension').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$7')\n",
    "\n",
    "y = df_mean_answer_step_by_hd_add_cp8.groupby('hidden_dimension').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_hd_add_cp8.groupby('hidden_dimension').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp8')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$8')\n",
    "\n",
    "y = df_mean_answer_step_by_hd_add_cp9.groupby('hidden_dimension').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_hd_add_cp9.groupby('hidden_dimension').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp9')\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$9')\n",
    "\n",
    "plt.legend(ncol=3, loc='upper left', fontsize=15)\n",
    "\n",
    "save_fig('mean_answer_step_by_hd-add', 'iccm2019')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2. Mean Answer Step By Hidden Dimension: Subtraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mean_answer_step_by_hd_sub_cp7 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_hd-sub-cp7.csv'))\n",
    "df_mean_answer_step_by_hd_sub_cp8 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_hd-sub-cp8.csv'))\n",
    "df_mean_answer_step_by_hd_sub_cp9 = pd.read_csv(join(config.dir_result_statistics(), experiment_name, 'mean_answer_step_by_hd-sub-cp9.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df_mean_answer_step_by_hd_sub_cp7['hidden_dimension'].unique()\n",
    "x.sort()\n",
    "\n",
    "ylimit = 3\n",
    "\n",
    "plt.figure(figsize=(6, ylimit))\n",
    "plt.xlabel(r'Hidden dimension ($d_c$)', fontsize=font_size['xlabel'])\n",
    "plt.ylabel('Mean answer step', fontsize=font_size['ylabel'])\n",
    "plt.tick_params(axis='x', labelsize=font_size['xtick'])\n",
    "plt.tick_params(axis='y', labelsize=font_size['ytick'])\n",
    "plt.xticks(x)\n",
    "plt.yticks(np.arange(0,ylimit+0.1,1))\n",
    "plt.ylim(bottom=0, top=ylimit)  \n",
    "plt.grid(axis='y')\n",
    "\n",
    "markersize = 12\n",
    "\n",
    "y = df_mean_answer_step_by_hd_sub_cp7.groupby('hidden_dimension').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_hd_sub_cp7.groupby('hidden_dimension').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp7')\n",
    "plt.plot(x, y, color='b', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$7')\n",
    "\n",
    "y = df_mean_answer_step_by_hd_sub_cp8.groupby('hidden_dimension').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_hd_sub_cp8.groupby('hidden_dimension').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp8')\n",
    "plt.plot(x, y, color='g', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$8')\n",
    "\n",
    "y = df_mean_answer_step_by_hd_sub_cp9.groupby('hidden_dimension').mean()['mean_answer_step'].values\n",
    "e = df_mean_answer_step_by_hd_sub_cp9.groupby('hidden_dimension').std()['mean_answer_step'].values\n",
    "#plt.errorbar(x, y, e, fmt='r:o', ecolor='orange', capsize=3, label='cp9')\n",
    "plt.plot(x, y, color='#cc0000', linestyle=':', marker='.', markersize=markersize, label=r'$\\mathtt{\\theta}$9')\n",
    "\n",
    "plt.legend(ncol=3, loc='upper left', fontsize=15)\n",
    "\n",
    "save_fig('mean_answer_step_by_hd-sub', 'iccm2019')"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}