loss_exp.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from utilities import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sigmoid_01(x):\n",
    "    return 1 / (1 + np.exp(-50*(x-0.1))) + 0.1*x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "x = np.arange(0, 1, 0.001)\n",
    "y = sigmoid_01(x)\n",
    "plt.plot(x, y); plt.xlabel('mape'); plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "plt.xlim(0, 0.5)\n",
    "plt.plot(x, x, label='mape')\n",
    "plt.plot(x, y, label='smooth hit rate'); \n",
    "plt.plot(x, [i > 0.1 for i in x], label='hit rate', alpha=0.7)\n",
    "plt.xlabel('MAPE'); plt.ylabel('loss')\n",
    "plt.grid(); plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred = pd.read_csv('output/model-26-lgb-search-bins-lr-0.0005-350-cv.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>building_id</th>\n",
       "      <th>total_price</th>\n",
       "      <th>total_price_predict</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>jre1pJhcQj91Kdky</td>\n",
       "      <td>14215011.0</td>\n",
       "      <td>1.274179e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>jcbuA8q3KPH9SzpS</td>\n",
       "      <td>7642884.5</td>\n",
       "      <td>9.406962e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>EqWJpHmkXPyfHkB8</td>\n",
       "      <td>4807053.0</td>\n",
       "      <td>4.658309e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ihrMWGM8Nq99Uvu6</td>\n",
       "      <td>7493026.5</td>\n",
       "      <td>6.930064e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>jRhwzoVMFLFzxAX1</td>\n",
       "      <td>2846855.8</td>\n",
       "      <td>2.784716e+06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        building_id  total_price  total_price_predict\n",
       "0  jre1pJhcQj91Kdky   14215011.0         1.274179e+07\n",
       "1  jcbuA8q3KPH9SzpS    7642884.5         9.406962e+06\n",
       "2  EqWJpHmkXPyfHkB8    4807053.0         4.658309e+06\n",
       "3  ihrMWGM8Nq99Uvu6    7493026.5         6.930064e+06\n",
       "4  jRhwzoVMFLFzxAX1    2846855.8         2.784716e+06"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred['ape'] = (df_pred['total_price_predict'] - df_pred['total_price']) / df_pred['total_price']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "plt.hist(df_pred['ape'], bins=200)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "plt.hist(df_pred['ape'], bins=200)\n",
    "plt.yscale('log')\n",
    "plt.xlabel('ape'); plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtcAAAEKCAYAAADZxnkxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAGURJREFUeJzt3X+wZ3V93/HnCxBMayqLbCgCk13tJgbTcXXuANZMVVRYoePiFO3SqqslsyaFTjJNZ7LETrUmjJtOEyZOjCnRrZgYkaLWbSHSFXGsnaAsFsGFIlfAYbcruwqi1gkVfPeP72fxy3Lv3u+993Pv9/54Pma+c8/3cz7n3M/5cL5fXvezn3NOqgpJkiRJ83fMuBsgSZIkrRSGa0mSJKkTw7UkSZLUieFakiRJ6sRwLUmSJHViuJYkSZI6MVxLkiRJnRiuJUmSpE4M15IkSVInx427AUdz8skn17p168bdDEmSJK1wt99++3eqau1897Okw/W6devYs2fPuJshSZKkFS7Jt3rsx2khkiRJUieGa0mSJKkTw7UkSZLUieFakiRJ6sRwLUmSJHViuJYkSZI6MVxLkiRJnRiuJUmSpE4M15IkSVInS/oJjZKk0azbfsNTyw/uuHCMLZGk1c2Ra0mSJKkTR64laYUZHsUGR7IlaTE5ci1JkiR1YriWJEmSOjFcS5IkSZ0YriVJkqRODNeSJElSJ4ZrSZIkqZMZw3WSZyf5SpKvJdmb5N+18vVJvpxkMsknkhzfyk9o7yfb+nVD+7qild+b5PyFOihJkiRpHEYZuX4cOLeqXgJsBDYlOQf4feCqqvp7wKPApa3+pcCjrfyqVo8kZwJbgBcDm4A/SXJsz4ORJEmSxmnGcF0DP2xvn9VeBZwLXN/KrwEuasub23va+tckSSu/tqoer6oHgEngrC5HIUmSJC0BI825TnJskjuAg8Bu4JvA96rqiVZlH3BaWz4NeAigrX8MeN5w+RTbSJIkScveSOG6qp6sqo3A6QxGm1+0UA1Ksi3JniR7Dh06tFC/RpIkSepuVncLqarvAbcALwdOTHJcW3U6sL8t7wfOAGjrnwt8d7h8im2Gf8fVVTVRVRNr166dTfMkSZKksRrlbiFrk5zYln8GeB1wD4OQfXGrthX4TFve1d7T1n++qqqVb2l3E1kPbAC+0utAJEmSpHE7buYqnApc0+7scQxwXVX9tyR3A9cm+T3gfwEfbvU/DPx5kkngEQZ3CKGq9ia5DrgbeAK4rKqe7Hs4kiRJ0vjMGK6r6k7gpVOU388Ud/uoqr8B3jTNvq4Erpx9MyVJkqSlzyc0SpIkSZ0YriVJkqRODNeSJElSJ4ZrSZIkqRPDtSRJktSJ4VqSJEnqxHAtSZIkdWK4liRJkjoxXEuSJEmdGK4lSZKkTgzXkiRJUieGa0mSJKkTw7UkSZLUieFakiRJ6sRwLUmSJHViuJYkSZI6MVxLkiRJnRiuJUmSpE4M15IkSVInhmtJkiSpE8O1JEmS1InhWpIkSerEcC1JkiR1MmO4TnJGkluS3J1kb5LfaOXvSbI/yR3tdcHQNlckmUxyb5Lzh8o3tbLJJNsX5pAkSZKk8ThuhDpPAL9VVV9N8rPA7Ul2t3VXVdV/GK6c5ExgC/Bi4PnA55L8Qlv9AeB1wD7gtiS7quruHgciSZIkjduM4bqqDgAH2vIPktwDnHaUTTYD11bV48ADSSaBs9q6yaq6HyDJta2u4VqSJEkrwqzmXCdZB7wU+HIrujzJnUl2JlnTyk4DHhrabF8rm65ckiRJWhFGDtdJngN8EvjNqvo+8EHghcBGBiPbf9CjQUm2JdmTZM+hQ4d67FKSJElaFCOF6yTPYhCsP1ZVnwKoqoer6smq+gnwZ/x06sd+4IyhzU9vZdOVP01VXV1VE1U1sXbt2tkejyRJkjQ2o9wtJMCHgXuq6g+Hyk8dqvZG4OtteRewJckJSdYDG4CvALcBG5KsT3I8g4sed/U5DEmSJGn8RrlbyCuAtwJ3Jbmjlf0OcEmSjUABDwLvBKiqvUmuY3Ch4hPAZVX1JECSy4GbgGOBnVW1t+OxSJIkSWM1yt1CvgRkilU3HmWbK4Erpyi/8WjbSZIkScuZT2iUJEmSOjFcS5IkSZ0YriVJkqRORrmgUZK0BK3bfsO4myBJOoIj15IkSVInhmtJkiSpE8O1JEmS1InhWpIkSerEcC1JkiR1YriWJEmSOjFcS5IkSZ0YriVJkqRODNeSJElSJ4ZrSZIkqRPDtSRJktSJ4VqSJEnqxHAtSZIkdWK4liRJkjoxXEuSJEmdGK4lSZKkTo4bdwMkSQtr3fYbnlp+cMeFY2yJJK18jlxLkiRJnRiuJUmSpE6cFiJJy8Tw9A5J0tI048h1kjOS3JLk7iR7k/xGKz8pye4k97Wfa1p5krw/yWSSO5O8bGhfW1v9+5JsXbjDkiRJkhbfKNNCngB+q6rOBM4BLktyJrAduLmqNgA3t/cArwc2tNc24IMwCOPAu4GzgbOAdx8O5JIkSdJKMGO4rqoDVfXVtvwD4B7gNGAzcE2rdg1wUVveDHy0Bm4FTkxyKnA+sLuqHqmqR4HdwKauRyNJkiSN0awuaEyyDngp8GXglKo60FZ9GzilLZ8GPDS02b5WNl35kb9jW5I9SfYcOnRoNs2TJEmSxmrkcJ3kOcAngd+squ8Pr6uqAqpHg6rq6qqaqKqJtWvX9tilJEmStChGCtdJnsUgWH+sqj7Vih9u0z1oPw+28v3AGUObn97KpiuXJEmSVoRR7hYS4MPAPVX1h0OrdgGH7/ixFfjMUPnb2l1DzgEea9NHbgLOS7KmXch4XiuTJEmSVoRR7nP9CuCtwF1J7mhlvwPsAK5LcinwLeDNbd2NwAXAJPAj4B0AVfVIkt8Fbmv13ltVj3Q5CknSSHwUuiQtrBnDdVV9Ccg0q18zRf0CLptmXzuBnbNpoCStZj44RpKWFx9/LkmSJHViuJYkSZI6MVxLkiRJnRiuJUmSpE4M15IkSVInhmtJkiSpE8O1JEmS1MkoD5GRJC0i720tScuXI9eSJElSJ4ZrSZIkqRPDtSRJktSJc64laZU6cm73gzsuHFNLJGnlcORakiRJ6sRwLUmSJHViuJYkSZI6cc61JAl4+hxs519L0twYriVpzHxojCStHE4LkSRJkjoxXEuSJEmdGK4lSZKkTgzXkiRJUieGa0mSJKkTw7UkSZLUyYzhOsnOJAeTfH2o7D1J9ie5o70uGFp3RZLJJPcmOX+ofFMrm0yyvf+hSJIkSeM1ysj1R4BNU5RfVVUb2+tGgCRnAluAF7dt/iTJsUmOBT4AvB44E7ik1ZUkSZJWjBkfIlNVX0yybsT9bQaurarHgQeSTAJntXWTVXU/QJJrW927Z91iSVoBfHCMJK1M85lzfXmSO9u0kTWt7DTgoaE6+1rZdOWSJEnSijHXcP1B4IXARuAA8Ae9GpRkW5I9SfYcOnSo124lSbOwbvsNT70kSaObU7iuqoer6smq+gnwZ/x06sd+4Iyhqqe3sunKp9r31VU1UVUTa9eunUvzJEmSpLGYU7hOcurQ2zcCh+8ksgvYkuSEJOuBDcBXgNuADUnWJzmewUWPu+bebEmSJGnpmfGCxiQfB14FnJxkH/Bu4FVJNgIFPAi8E6Cq9ia5jsGFik8Al1XVk20/lwM3AccCO6tqb/ejkSRJksZolLuFXDJF8YePUv9K4Mopym8EbpxV6yRphXDusiStDj6hUZIkSerEcC1JkiR1MuO0EEnS3DgVRJJWH8O1JOmojvwj4cEdF46pJZK09DktRJIkSerEcC1JkiR1YriWJEmSOjFcS5IkSZ0YriVJkqRODNeSJElSJ4ZrSZIkqRPvcy1JHfngGEla3Ry5liRJkjoxXEuSJEmdGK4lSZKkTpxzLUmaleF55Q/uuHCMLZGkpceRa0mSJKkTw7UkSZLUidNCJGkevPWeJGmYI9eSJElSJ45cS5LmzIsbJenpHLmWJEmSOjFcS5IkSZ3MOC0kyU7gHwEHq+qXW9lJwCeAdcCDwJur6tEkAf4IuAD4EfD2qvpq22Yr8G/abn+vqq7peyiSpHE68uJOp4lIWo1GmXP9EeCPgY8OlW0Hbq6qHUm2t/e/Dbwe2NBeZwMfBM5uYfzdwARQwO1JdlXVo70ORJIWi3cIkSRNZ8ZpIVX1ReCRI4o3A4dHnq8BLhoq/2gN3AqcmORU4Hxgd1U90gL1bmBTjwOQJEmSloq5zrk+paoOtOVvA6e05dOAh4bq7Wtl05VLkiRJK8a8L2isqmIw1aOLJNuS7Emy59ChQ712K0mSJC24uYbrh9t0D9rPg618P3DGUL3TW9l05c9QVVdX1URVTaxdu3aOzZMkSZIW31zD9S5ga1veCnxmqPxtGTgHeKxNH7kJOC/JmiRrgPNamSRphVq3/YanXpK0WoxyK76PA68CTk6yj8FdP3YA1yW5FPgW8OZW/UYGt+GbZHArvncAVNUjSX4XuK3Ve29VHXmRpCRJkrSszRiuq+qSaVa9Zoq6BVw2zX52Ajtn1TpJkiRpGfEJjZIkSVInhmtJkiSpE8O1JEmS1Mkojz+XpFXPO15IkkbhyLUkSZLUiSPXkqRFdeS/Ajy448IxtUSS+jNcS9IUnAbSl/0pabVwWogkSZLUieFakiRJ6sRwLUmSJHXinGtJapwXLEmaL0euJUmSpE4M15IkSVInhmtJkiSpE+dcS5LGaniuuw+UkbTcOXItSZIkdeLItSRpyXAUW9JyZ7iWtKp5+z1JUk9OC5EkSZI6MVxLkiRJnTgtRNKq4jSQ5ePI/1bOwZa0HDhyLUmSJHViuJYkSZI6cVqIpBXPqSArg7fpk7QczGvkOsmDSe5KckeSPa3spCS7k9zXfq5p5Uny/iSTSe5M8rIeByBJkiQtFT2mhby6qjZW1UR7vx24uao2ADe39wCvBza01zbggx1+tyRJkrRkLMSc683ANW35GuCiofKP1sCtwIlJTl2A3y9JkiSNxXzDdQH/PcntSba1slOq6kBb/jZwSls+DXhoaNt9rexpkmxLsifJnkOHDs2zeZIkSdLime8Fjb9SVfuT/BywO8n/Hl5ZVZWkZrPDqroauBpgYmJiVttKklYnL3aUtFTMK1xX1f7282CSTwNnAQ8nObWqDrRpHwdb9f3AGUObn97KJGnevCOIJGkpmHO4TvK3gWOq6gdt+TzgvcAuYCuwo/38TNtkF3B5kmuBs4HHhqaPSJI0Mv+YkrRUzWfk+hTg00kO7+cvq+qzSW4DrktyKfAt4M2t/o3ABcAk8CPgHfP43ZIkSdKSM+dwXVX3Ay+Zovy7wGumKC/gsrn+PkmSJGmp8wmNkpYtpwZoKl7cKGmcFuI+15IkSdKqZLiWJEmSOjFcS5IkSZ0451rSkua8as3H0c4f52NLWgiOXEuSJEmdGK4lSZKkTpwWImnJcSqIFoO37JO0EAzXkqRVz7nZknoxXEtaEhytliStBIZrSWNhmNZy4ai2pNnwgkZJkiSpE0euJUmaIy+KlHQkw7WkReNUEEnSSme4lrSgDNRaLRzFlgSGa0mdGaalZ34ODNvS6mG4ljRvBmppdI5wSytbqmrcbZjWxMRE7dmzZ9zNkHQEw7S0MAzb0vgkub2qJua7H0euJUlaIkb9w9UQLi1dhmtJI3G0Wlo6pvs8Grql8TNcS3qKAVqSpPkxXEsryKgXShmipZVprp/tUb8vHBmXZuYFjdIyYzCWtBQYtLXSLNsLGpNsAv4IOBb4UFXtWOw2SMuNgVrSUnO0e3l7YaZWs0UduU5yLPAN4HXAPuA24JKqunuq+o5ca7may/9YDNCS9FPzDeujhv/ZBHynyKxsvUauFztcvxx4T1Wd395fAVBV75uqvuFa43C0L08DsCRpJtMF77kGfu8OsziWa7i+GNhUVb/a3r8VOLuqLp+qvuF64czm0bzTfeDn+k+CjtZKkqSlYDiTrNhwnWQbsK29/WXg64vWwJXvZOA7427ECmJ/9mV/9mNf9mV/9mV/9mNf9vWLVfWz893JYl/QuB84Y+j96a3sKVV1NXA1QJI9Pf6C0ID92Zf92Zf92Y992Zf92Zf92Y992VeSLtMljumxk1m4DdiQZH2S44EtwK5FboMkSZK0IBZ15LqqnkhyOXATg1vx7ayqvYvZBkmSJGmhLPp9rqvqRuDGEatfvZBtWYXsz77sz77sz37sy77sz77sz37sy7669OeSfkKjJEmStJws9pxrSZIkacUae7hO8qYke5P8JMm0V7wm2ZTk3iSTSbYPla9P8uVW/ol2oeSqleSkJLuT3Nd+rpmizquT3DH0+pskF7V1H0nywNC6jYt/FEvHKP3Z6j051Ge7hso9P5sRz82NSf66fSfcmeSfDK3z3GT678Kh9Se0c22ynXvrhtZd0crvTXL+YrZ7KRqhL/9VkrvbuXhzkp8fWjflZ341G6E/357k0FC//erQuq3tu+G+JFsXt+VL0wj9edVQX34jyfeG1nl+DkmyM8nBJFPe3jkD7299fWeSlw2tm/25WVVjfQG/BPwi8AVgYpo6xwLfBF4AHA98DTizrbsO2NKW/xT49XEf05j7898D29vyduD3Z6h/EvAI8Lfa+48AF4/7OJbKa9T+BH44Tbnn5yz6EvgFYENbfj5wADixvV/15+bRvguH6vwL4E/b8hbgE235zFb/BGB928+x4z6mJd6Xrx76bvz1w33Z3k/5mV+trxH78+3AH0+x7UnA/e3nmra8ZtzHtNT784j6/5LBTSIOv/f8fHr//EPgZcDXp1l/AfBXQIBzgC+38jmdm2Mfua6qe6rq3hmqnQVMVtX9VfX/gGuBzUkCnAtc3+pdA1y0cK1dFjYz6AcYrT8uBv6qqn60oK1avmbbn0/x/HyGGfuyqr5RVfe15f8DHATWLloLl74pvwuPqDPcz9cDr2nn4mbg2qp6vKoeACbb/larGfuyqm4Z+m68lcGzGTS1Uc7N6ZwP7K6qR6rqUWA3sGmB2rlczLY/LwE+vigtW4aq6osMBhKnsxn4aA3cCpyY5FTmeG6OPVyP6DTgoaH3+1rZ84DvVdUTR5SvZqdU1YG2/G3glBnqb+GZH8gr2z+LXJXkhO4tXF5G7c9nJ9mT5NbDU2zw/DzSrM7NJGcxGLH55lDxaj83p/sunLJOO/ceY3AujrLtajLb/riUwcjWYVN95lezUfvzH7fP8PVJDj9UznPzmUbukzZdaT3w+aFiz8/Zma6/53RuLsqt+JJ8Dvi7U6x6V1V9ZjHasJIcrT+H31RVJZn2djDtr7K/z+C+44ddwSD4HM/gljS/Dbx3vm1eyjr1589X1f4kLwA+n+QuBqFmVel8bv45sLWqftKKV925qaUhyVuACeCVQ8XP+MxX1Ten3oOa/wp8vKoeT/JOBv/Ccu6Y27QSbAGur6onh8o8P8doUcJ1Vb12nruY7rHp32UwdH9cG6F5xuPUV6Kj9WeSh5OcWlUHWkA5eJRdvRn4dFX9eGjfh0cWH0/yn4B/3aXRS1iP/qyq/e3n/Um+ALwU+CSr7Pzs0ZdJ/g5wA4M/vm8d2veqOzenMN134VR19iU5Dngug+/KUbZdTUbqjySvZfDH4Sur6vHD5dN85ldzeJmxP6vqu0NvP8TgOozD277qiG2/0L2Fy8tsPq9bgMuGCzw/Z226/p7TublcpoVM+dj0Gsw2v4XBvGGArcBqHwnfxaAfYOb+eMYcrRZ6Ds8XvgiY8sraVWTG/kyy5vAUhSQnA68A7vb8fIZR+vJ44NMM5r5df8Q6z81pvguPqDPczxcDn2/n4i5gSwZ3E1kPbAC+skjtXopm7MskLwX+I/CGqjo4VD7lZ37RWr40jdKfpw69fQNwT1u+CTiv9esa4Dye/i+qq9Eon3WSvIjBhXZ/PVTm+Tl7u4C3tbuGnAM81gZ05nZuLubVmtNcoflGBnNYHgceBm5q5c8HbhyqdwHwDQZ/eb1rqPwFDP4HMQn8Z+CEcR/TmPvzecDNwH3A54CTWvkE8KGheusY/EV2zBHbfx64i0Fw+QvgOeM+pqXen8A/aH32tfbz0qHtPT9n15dvAX4M3DH02tjWeW7W1N+FDKbHvKEtP7uda5Pt3HvB0LbvatvdC7x+3Mcy7tcIffm59v+lw+firlY+7Wd+Nb9G6M/3AXtbv90CvGho23/eztlJ4B3jPpal8JqpP9v79wA7jtjO8/OZfflxBnef+jGDzHkp8GvAr7X1AT7Q+vouhu5eN5dz0yc0SpIkSZ0sl2khkiRJ0pJnuJYkSZI6MVxLkiRJnRiuJUmSpE4M15IkSVInhmtJkiSpE8O1JEmS1InhWpKWoST/JcntSfYm2dbKfpjkqlZ2c5K1rfyFST7b6v+P9lQ3SdIC8CEykrQMJTmpqh5J8jMMHpX8SuA7wFuq6mNJ/i3wc1V1eZKbGTyJ7L4kZwPvq6pzx9h8SVqxDNeStAwleQ/wxvZ2HXA+8D+BE6rqiSQvAD4F/ApwiMEjzw87oap+afFaK0mrx3HjboAkaXaSvAp4LfDyqvpRki8Az56iajGY/ve9qtq4eC2UpNXLOdeStPw8F3i0BesXAee08mOAi9vyPwW+VFXfBx5I8iaADLxk0VssSauE4VqSlp/PAscluQfYAdzayv8vcFaSrwPnAu9t5f8MuDTJ14C9wOZFbq8krRrOuZakFSLJD6vqOeNuhyStZo5cS5IkSZ04ci1JkiR14si1JEmS1InhWpIkSerEcC1JkiR1YriWJEmSOjFcS5IkSZ0YriVJkqRO/j8N7dfRUblksQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "plt.hist(df_pred['ape'], bins=np.arange(-1,1,0.01))\n",
    "plt.xlim(-1, 1)\n",
    "plt.xlabel('ape'); plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 13"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred = pd.read_csv('output/model-13-predict-keras-he_uni-cv.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>building_id</th>\n",
       "      <th>total_price</th>\n",
       "      <th>total_price_predict</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>jre1pJhcQj91Kdky</td>\n",
       "      <td>14215011.0</td>\n",
       "      <td>12871026.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>jcbuA8q3KPH9SzpS</td>\n",
       "      <td>7642884.5</td>\n",
       "      <td>6722597.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>EqWJpHmkXPyfHkB8</td>\n",
       "      <td>4807053.0</td>\n",
       "      <td>4915826.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ihrMWGM8Nq99Uvu6</td>\n",
       "      <td>7493026.5</td>\n",
       "      <td>7121956.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>jRhwzoVMFLFzxAX1</td>\n",
       "      <td>2846855.8</td>\n",
       "      <td>2657166.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        building_id  total_price  total_price_predict\n",
       "0  jre1pJhcQj91Kdky   14215011.0           12871026.0\n",
       "1  jcbuA8q3KPH9SzpS    7642884.5            6722597.0\n",
       "2  EqWJpHmkXPyfHkB8    4807053.0            4915826.5\n",
       "3  ihrMWGM8Nq99Uvu6    7493026.5            7121956.5\n",
       "4  jRhwzoVMFLFzxAX1    2846855.8            2657166.2"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cal_score_smooth(y_true, y_pred):\n",
    "    pe = np.abs(y_pred - y_true) / y_true\n",
    "    return 100*np.mean(sigmoid_01(pe))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sigmoid_01(x):\n",
    "    return 1 / (1 + np.exp(-50*(x-0.1))) + 0.1*x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "48.101514640571175"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cal_score_smooth(df_pred['total_price'], df_pred['total_price_predict'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred['ape'] = (df_pred['total_price_predict'] - df_pred['total_price']) / df_pred['total_price']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtUAAAD8CAYAAAC4j1i8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAFMhJREFUeJzt3W+sZVd5H+DfGw8QIAUbPHGpx2TcMmpkUAEzMo6IKmq39hgQ5gOhRrSeUherwqikokptItUKBAnUKiSoBMnCLnZEMRaBYhETZ2SI0n6w8Zg/BuNQTw3EYxk8YGOSokBN3n44y8lhuOO5M+vO3Hud55GOzt7vXvvcdZbWzP3NnnX2qe4OAABw9H5mvTsAAACbnVANAACThGoAAJgkVAMAwCShGgAAJgnVAAAwSagGAIBJQjUAAEwSqgEAYNKW9e7A0Tr55JN7+/bt690NAACewO64447vdPfWw7XbtKF6+/bt2bt373p3AwCAJ7Cq+uZq2ln+AQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJM27TcqbgTbL/+Dv97+xrtfuY49AQBgPblSDQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADBpVaG6qr5RVV+uqi9W1d5Re1ZV7amqe8bzSaNeVfW+qtpXVXdW1ZlLr7N7tL+nqnYv1V8yXn/fOLfW+o0CAMCxciRXqv9Jd7+ou3eO/cuT3NLdO5LcMvaT5IIkO8bj0iQfSBYhPMmVSV6a5KwkVz4WxEebNy2dt+uo3xEAABxnM8s/Lkxy7di+NslrlurX9cKtSU6squckOT/Jnu5+qLsfTrInya5x7BndfWt3d5Lrll4LAAA2vNWG6k7yR1V1R1VdOmqndPcDY/tbSU4Z26cmuW/p3P2j9nj1/SvUAQBgU9iyyna/3N33V9XPJ9lTVX+6fLC7u6p67bv3k0agvzRJnvvc5x7rHwcAAKuyqivV3X3/eH4wySeyWBP97bF0I+P5wdH8/iSnLZ2+bdQer75thfpK/biqu3d2986tW7eupusAAHDMHTZUV9XTq+rvPLad5LwkX0lyY5LH7uCxO8knx/aNSS4edwE5O8kjY5nIzUnOq6qTxgcUz0ty8zj2/ao6e9z14+Kl1wIAgA1vNcs/TknyiXGXuy1J/nt3/2FV3Z7khqq6JMk3k7xutL8pySuS7EvygyRvTJLufqiq3pnk9tHuHd390Nh+c5IPJXlqkk+PBwAAbAqHDdXdfW+SF65Q/26Sc1eod5LLDvFa1yS5ZoX63iQvWEV/AQBgw/GNigAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADBJqAYAgElCNQAATBKqAQBgklANAACThGoAAJgkVAMAwCShGgAAJgnVAAAwSagGAIBJQjUAAEwSqgEAYJJQDQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJNWHaqr6oSq+kJVfWrsn15Vt1XVvqr6aFU9edSfMvb3jePbl17jilH/WlWdv1TfNWr7qurytXt7AABw7B3Jleq3Jrl7af89Sd7b3c9L8nCSS0b9kiQPj/p7R7tU1RlJLkry/CS7kvzuCOonJHl/kguSnJHk9aMtAABsCqsK1VW1Lckrk3xw7FeSc5J8bDS5NslrxvaFYz/j+Lmj/YVJru/uH3b315PsS3LWeOzr7nu7+0dJrh9tAQBgU1jtlerfTvJrSf5q7D87yfe6+9Gxvz/JqWP71CT3Jck4/sho/9f1g845VB0AADaFw4bqqnpVkge7+47j0J/D9eXSqtpbVXsPHDiw3t0BAIAkq7tS/bIkr66qb2SxNOOcJL+T5MSq2jLabEty/9i+P8lpSTKOPzPJd5frB51zqPpP6e6runtnd+/cunXrKroOAADH3mFDdXdf0d3bunt7Fh80/Ex3vyHJZ5O8djTbneSTY/vGsZ9x/DPd3aN+0bg7yOlJdiT5XJLbk+wYdxN58vgZN67JuwMAgONgy+GbHNJ/THJ9Vf1mki8kuXrUr07ye1W1L8lDWYTkdPddVXVDkq8meTTJZd394ySpqrckuTnJCUmu6e67JvoFAADH1RGF6u7+4yR/PLbvzeLOHQe3+cskv3KI89+V5F0r1G9KctOR9AUAADYK36gIAACThGoAAJgkVAMAwCShGgAAJgnVAAAwSagGAIBJQjUAAEwSqgEAYJJQDQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADBJqAYAgElCNQAATBKqAQBgklANAACThGoAAJgkVAMAwCShGgAAJgnVAAAw6bChuqp+tqo+V1Vfqqq7quo3Rv30qrqtqvZV1Uer6smj/pSxv28c3770WleM+teq6vyl+q5R21dVl6/92wQAgGNnNVeqf5jknO5+YZIXJdlVVWcneU+S93b385I8nOSS0f6SJA+P+ntHu1TVGUkuSvL8JLuS/G5VnVBVJyR5f5ILkpyR5PWjLQAAbAqHDdW98Bdj90nj0UnOSfKxUb82yWvG9oVjP+P4uVVVo359d/+wu7+eZF+Ss8ZjX3ff290/SnL9aAsAAJvCqtZUjyvKX0zyYJI9Sf5Pku9196Ojyf4kp47tU5PclyTj+CNJnr1cP+icQ9VX6selVbW3qvYeOHBgNV0HAIBjblWhurt/3N0vSrItiyvLv3hMe3XoflzV3Tu7e+fWrVvXowsAAPBTjujuH939vSSfTfJLSU6sqi3j0LYk94/t+5OcliTj+DOTfHe5ftA5h6oDAMCmsJq7f2ytqhPH9lOT/LMkd2cRrl87mu1O8smxfePYzzj+me7uUb9o3B3k9CQ7knwuye1Jdoy7iTw5iw8z3rgWbw4AAI6HLYdvkuckuXbcpeNnktzQ3Z+qqq8mub6qfjPJF5JcPdpfneT3qmpfkoeyCMnp7ruq6oYkX03yaJLLuvvHSVJVb0lyc5ITklzT3Xet2TsEAIBj7LChurvvTPLiFer3ZrG++uD6Xyb5lUO81ruSvGuF+k1JblpFfwEAYMPxjYoAADBJqAYAgElCNQAATBKqAQBgklANAACThGoAAJgkVAMAwCShGgAAJgnVAAAwSagGAIBJQjUAAEwSqgEAYJJQDQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADBJqAYAgElCNQAATBKqAQBgklANAACTDhuqq+q0qvpsVX21qu6qqreO+rOqak9V3TOeTxr1qqr3VdW+qrqzqs5ceq3do/09VbV7qf6SqvryOOd9VVXH4s0CAMCxsJor1Y8meVt3n5Hk7CSXVdUZSS5Pckt370hyy9hPkguS7BiPS5N8IFmE8CRXJnlpkrOSXPlYEB9t3rR03q75twYAAMfHYUN1dz/Q3Z8f23+e5O4kpya5MMm1o9m1SV4zti9Mcl0v3JrkxKp6TpLzk+zp7oe6++Eke5LsGsee0d23dncnuW7ptQAAYMM7ojXVVbU9yYuT3JbklO5+YBz6VpJTxvapSe5bOm3/qD1eff8K9ZV+/qVVtbeq9h44cOBIug4AAMfMqkN1Vf1ckt9P8qvd/f3lY+MKc69x335Kd1/V3Tu7e+fWrVuP9Y8DAIBVWVWorqonZRGoP9zdHx/lb4+lGxnPD476/UlOWzp926g9Xn3bCnUAANgUVnP3j0pydZK7u/u3lg7dmOSxO3jsTvLJpfrF4y4gZyd5ZCwTuTnJeVV10viA4nlJbh7Hvl9VZ4+fdfHSawEAwIa3ZRVtXpbkXyb5clV9cdTenuTdSW6oqkuSfDPJ68axm5K8Ism+JD9I8sYk6e6HquqdSW4f7d7R3Q+N7Tcn+VCSpyb59HgAAMCmcNhQ3d3/K8mh7ht97grtO8llh3ita5Jcs0J9b5IXHK4vAACwEflGRQAAmCRUAwDAJKEaAAAmCdUAADBJqAYAgElCNQAATBKqAQBgklANAACThGoAAJgkVAMAwCShGgAAJgnVAAAwSagGAIBJQjUAAEwSqgEAYJJQDQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADDpsKG6qq6pqger6itLtWdV1Z6qumc8nzTqVVXvq6p9VXVnVZ25dM7u0f6eqtq9VH9JVX15nPO+qqq1fpMAAHAsreZK9YeS7DqodnmSW7p7R5Jbxn6SXJBkx3hcmuQDySKEJ7kyyUuTnJXkyseC+GjzpqXzDv5ZAACwoR02VHf3nyR56KDyhUmuHdvXJnnNUv26Xrg1yYlV9Zwk5yfZ090PdffDSfYk2TWOPaO7b+3uTnLd0msBAMCmcLRrqk/p7gfG9reSnDK2T01y31K7/aP2ePX9K9QBAGDTmP6g4rjC3GvQl8Oqqkuram9V7T1w4MDx+JEAAHBYRxuqvz2WbmQ8Pzjq9yc5bandtlF7vPq2Feor6u6runtnd+/cunXrUXYdAADW1tGG6huTPHYHj91JPrlUv3jcBeTsJI+MZSI3Jzmvqk4aH1A8L8nN49j3q+rscdePi5deCwAANoUth2tQVR9J8vIkJ1fV/izu4vHuJDdU1SVJvpnkdaP5TUlekWRfkh8keWOSdPdDVfXOJLePdu/o7sc+/PjmLO4w8tQknx4PAADYNA4bqrv79Yc4dO4KbTvJZYd4nWuSXLNCfW+SFxyuHwAAsFH5RkUAAJgkVAMAwCShGgAAJgnVAAAwSagGAIBJQjUAAEwSqgEAYJJQDQAAk4RqAACYJFQDAMAkoRoAACYJ1QAAMEmoBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADBpy3p3YLPZfvkfrHcXAADYYFypBgCASUI1AABMEqoBAGCSUA0AAJOEagAAmOTuH2tk+a4g33j3K9exJwAAHG+uVAMAwCShGgAAJm2YUF1Vu6rqa1W1r6ouX+/+AADAam2IUF1VJyR5f5ILkpyR5PVVdcb69goAAFZnQ4TqJGcl2dfd93b3j5Jcn+TCde4TAACsyka5+8epSe5b2t+f5KXr1Jdp7gQCAPC3y0YJ1atSVZcmuXTs/kVVfW09+zOcnOQ7hzpY7zmOPXnieNwx5agZ17VnTNeeMV17xnTtGdO1t5HH9BdW02ijhOr7k5y2tL9t1H5Cd1+V5Krj1anVqKq93b1zvfvxRGJMjw3juvaM6dozpmvPmK49Y7r2nghjulHWVN+eZEdVnV5VT05yUZIb17lPAACwKhviSnV3P1pVb0lyc5ITklzT3Xetc7cAAGBVNkSoTpLuvinJTevdj6OwoZajPEEY02PDuK49Y7r2jOnaM6Zrz5iuvU0/ptXd690HAADY1DbKmmoAANi0hOoJvlp9XlWdVlWfraqvVtVdVfXWUX9WVe2pqnvG80nr3dfNpqpOqKovVNWnxv7pVXXbmK8fHR8KZpWq6sSq+lhV/WlV3V1Vv2Sezqmqfz/+3H+lqj5SVT9rnh65qrqmqh6sqq8s1Vacm7XwvjG+d1bVmevX843rEGP6n8ef/zur6hNVdeLSsSvGmH6tqs5fn15vbCuN6dKxt1VVV9XJY39TzlOh+ij5avU182iSt3X3GUnOTnLZGMfLk9zS3TuS3DL2OTJvTXL30v57kry3u5+X5OEkl6xLrzav30nyh939i0lemMXYmqdHqapOTfLvkuzs7hdk8SH1i2KeHo0PJdl1UO1Qc/OCJDvG49IkHzhOfdxsPpSfHtM9SV7Q3f8oyf9OckWSjN9ZFyV5/jjnd0dG4Cd9KD89pqmq05Kcl+TPlsqbcp4K1UfPV6uvge5+oLs/P7b/PIugcmoWY3ntaHZtktesTw83p6raluSVST449ivJOUk+NpoY0yNQVc9M8o+TXJ0k3f2j7v5ezNNZW5I8taq2JHlakgdinh6x7v6TJA8dVD7U3LwwyXW9cGuSE6vqOcenp5vHSmPa3X/U3Y+O3Vuz+E6NZDGm13f3D7v760n2ZZERWHKIeZok703ya0mWP+S3KeepUH30Vvpq9VPXqS9PCFW1PcmLk9yW5JTufmAc+laSU9apW5vVb2fxl9Rfjf1nJ/ne0i8E8/XInJ7kQJL/NpbUfLCqnh7z9Kh19/1J/ksWV6ceSPJIkjtinq6VQ81Nv7vWxr9O8umxbUyPUlVdmOT+7v7SQYc25ZgK1WwIVfVzSX4/ya929/eXj/XiFjVuU7NKVfWqJA929x3r3ZcnkC1Jzkzyge5+cZL/m4OWepinR2as8b0wi3+w/L0kT88K/zXMPHNzbVXVr2exdPHD692Xzayqnpbk7Un+03r3Za0I1UdvVV+tzuFV1ZOyCNQf7u6Pj/K3H/uvnvH84Hr1bxN6WZJXV9U3sliWdE4W64FPHP/NnpivR2p/kv3dfdvY/1gWIds8PXr/NMnXu/tAd/+/JB/PYu6ap2vjUHPT764JVfWvkrwqyRv6b+5JbEyPzj/I4h/VXxq/r7Yl+XxV/d1s0jEVqo+er1ZfA2Ot79VJ7u7u31o6dGOS3WN7d5JPHu++bVbdfUV3b+vu7VnMy8909xuSfDbJa0czY3oEuvtbSe6rqn84Sucm+WrM0xl/luTsqnra+HvgsTE1T9fGoebmjUkuHndXODvJI0vLRHgcVbUri2V1r+7uHywdujHJRVX1lKo6PYsP131uPfq4mXT3l7v757t7+/h9tT/JmePv2005T335y4SqekUWa1cf+2r1d61zlzadqvrlJP8zyZfzN+t/357Fuuobkjw3yTeTvK67V/qAA4+jql6e5D9096uq6u9nceX6WUm+kORfdPcP17N/m0lVvSiLD34+Ocm9Sd6YxYUJ8/QoVdVvJPnnWfxX+heS/Jss1k2ap0egqj6S5OVJTk7y7SRXJvkfWWFujn/A/Ncsltr8IMkbu3vvevR7IzvEmF6R5ClJvjua3drd/3a0//Us1lk/msUyxk8f/Jp/2600pt199dLxb2RxN6DvbNZ5KlQDAMAkyz8AAGCSUA0AAJOEagAAmCRUAwDAJKEaAAAmCdUAADBJqAYAgElCNQAATPr/98nK5GRGkSsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "plt.hist(df_pred['ape'], bins=200)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "plt.hist(df_pred['ape'], bins=200)\n",
    "plt.yscale('log')\n",
    "plt.xlabel('ape'); plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "plt.hist(df_pred['ape'], bins=np.arange(-1,1,0.01))\n",
    "plt.xlim(-1, 1)\n",
    "plt.xlabel('ape'); plt.show()"
   ]
  },
  {
   "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": "3.6.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "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
}