{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "FhGuhbZ6M5tl"
   },
   "source": [
    "##### Copyright 2018 The TensorFlow Authors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "cellView": "form",
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:01.154918Z",
     "iopub.status.busy": "2022-08-09T01:44:01.154377Z",
     "iopub.status.idle": "2022-08-09T01:44:01.158343Z",
     "shell.execute_reply": "2022-08-09T01:44:01.157744Z"
    },
    "id": "AwOEIRJC6Une"
   },
   "outputs": [],
   "source": [
    "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
    "# you may not use this file except in compliance with the License.\n",
    "# You may obtain a copy of the License at\n",
    "#\n",
    "# https://www.apache.org/licenses/LICENSE-2.0\n",
    "#\n",
    "# Unless required by applicable law or agreed to in writing, software\n",
    "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
    "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
    "# See the License for the specific language governing permissions and\n",
    "# limitations under the License."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "cellView": "form",
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:01.161509Z",
     "iopub.status.busy": "2022-08-09T01:44:01.161107Z",
     "iopub.status.idle": "2022-08-09T01:44:01.164422Z",
     "shell.execute_reply": "2022-08-09T01:44:01.163909Z"
    },
    "id": "KyPEtTqk6VdG"
   },
   "outputs": [],
   "source": [
    "#@title MIT License\n",
    "#\n",
    "# Copyright (c) 2017 François Chollet\n",
    "#\n",
    "# Permission is hereby granted, free of charge, to any person obtaining a\n",
    "# copy of this software and associated documentation files (the \"Software\"),\n",
    "# to deal in the Software without restriction, including without limitation\n",
    "# the rights to use, copy, modify, merge, publish, distribute, sublicense,\n",
    "# and/or sell copies of the Software, and to permit persons to whom the\n",
    "# Software is furnished to do so, subject to the following conditions:\n",
    "#\n",
    "# The above copyright notice and this permission notice shall be included in\n",
    "# all copies or substantial portions of the Software.\n",
    "#\n",
    "# THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n",
    "# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n",
    "# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL\n",
    "# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n",
    "# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING\n",
    "# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER\n",
    "# DEALINGS IN THE SOFTWARE."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EIdT9iu_Z4Rb"
   },
   "source": [
    "# 回帰：燃費を予測する"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bBIlTPscrIT9"
   },
   "source": [
    "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
    "  <td><a target=\"_blank\" href=\"https://www.tensorflow.org/tutorials/keras/regression\"><img src=\"https://www.tensorflow.org/images/tf_logo_32px.png\">TensorFlow.org で表示</a></td>\n",
    "  <td>     <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/docs-l10n/blob/master/site/ja/tutorials/keras/regression.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\">Google Colab で実行</a> </td>\n",
    "  <td><a target=\"_blank\" href=\"https://github.com/tensorflow/docs-l10n/blob/master/site/ja/tutorials/keras/regression.ipynb\">     <img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\">     GitHubでソースを表示</a></td>\n",
    "  <td>     <a href=\"https://storage.googleapis.com/tensorflow_docs/docs-l10n/site/ja/tutorials/keras/regression.ipynb\"><img src=\"https://www.tensorflow.org/images/download_logo_32px.png\">ノートブックをダウンロード</a> </td>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "AHp3M9ZmrIxj"
   },
   "source": [
    "回帰問題では、価格や確率といった連続的な値の出力を予測することが目的となります。これは、分類問題の目的が、（たとえば、写真にリンゴが写っているかオレンジが写っているかといった）離散的なラベルを予測することであるのとは対照的です。\n",
    "\n",
    "このノートブックでは、古典的な [Auto MPG](https://archive.ics.uci.edu/ml/datasets/auto+mpg) データセットを使用し、1970 年代後半から 1980 年台初めの自動車の燃費を予測するモデルを構築します。この目的のため、モデルにはこの時期の多数の自動車の仕様を読み込ませます。仕様には、気筒数、排気量、馬力、重量などが含まれています。\n",
    "\n",
    "このサンプルでは`tf.keras` APIを使用しています。詳細は[このガイド](https://www.tensorflow.org/guide/keras)を参照してください。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:01.167876Z",
     "iopub.status.busy": "2022-08-09T01:44:01.167364Z",
     "iopub.status.idle": "2022-08-09T01:44:02.492703Z",
     "shell.execute_reply": "2022-08-09T01:44:02.491739Z"
    },
    "id": "moB4tpEHxKB3"
   },
   "outputs": [],
   "source": [
    "# Use seaborn for pairplot.\n",
    "!pip install -q seaborn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:02.496822Z",
     "iopub.status.busy": "2022-08-09T01:44:02.496584Z",
     "iopub.status.idle": "2022-08-09T01:44:03.698393Z",
     "shell.execute_reply": "2022-08-09T01:44:03.697752Z"
    },
    "id": "1rRo8oNqZ-Rj"
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "# Make NumPy printouts easier to read.\n",
    "np.set_printoptions(precision=3, suppress=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:03.702459Z",
     "iopub.status.busy": "2022-08-09T01:44:03.701806Z",
     "iopub.status.idle": "2022-08-09T01:44:05.202395Z",
     "shell.execute_reply": "2022-08-09T01:44:05.201724Z"
    },
    "id": "9xQKvCJ85kCQ"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-08-09 01:44:03.878370: E tensorflow/stream_executor/cuda/cuda_blas.cc:2981] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-08-09 01:44:04.562175: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvrtc.so.11.1: cannot open shared object file: No such file or directory\n",
      "2022-08-09 01:44:04.562514: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvrtc.so.11.1: cannot open shared object file: No such file or directory\n",
      "2022-08-09 01:44:04.562528: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.10.0-rc0\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "F_72b0LCNbjx"
   },
   "source": [
    "## Auto MPG データセット\n",
    "\n",
    "このデータセットは[UCI Machine Learning Repository](https://archive.ics.uci.edu/ml/)から入手可能です。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "gFh9ne3FZ-On"
   },
   "source": [
    "### データの取得\n",
    "\n",
    "まず、データセットをダウンロードします。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.206320Z",
     "iopub.status.busy": "2022-08-09T01:44:05.205501Z",
     "iopub.status.idle": "2022-08-09T01:44:05.547988Z",
     "shell.execute_reply": "2022-08-09T01:44:05.547229Z"
    },
    "id": "CiX2FI4gZtTt"
   },
   "outputs": [],
   "source": [
    "url = 'http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data'\n",
    "column_names = ['MPG', 'Cylinders', 'Displacement', 'Horsepower', 'Weight',\n",
    "                'Acceleration', 'Model Year', 'Origin']\n",
    "\n",
    "raw_dataset = pd.read_csv(url, names=column_names,\n",
    "                          na_values='?', comment='\\t',\n",
    "                          sep=' ', skipinitialspace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.552344Z",
     "iopub.status.busy": "2022-08-09T01:44:05.551707Z",
     "iopub.status.idle": "2022-08-09T01:44:05.569121Z",
     "shell.execute_reply": "2022-08-09T01:44:05.568528Z"
    },
    "id": "2oY3pMPagJrO"
   },
   "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>MPG</th>\n",
       "      <th>Cylinders</th>\n",
       "      <th>Displacement</th>\n",
       "      <th>Horsepower</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>Model Year</th>\n",
       "      <th>Origin</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>393</th>\n",
       "      <td>27.0</td>\n",
       "      <td>4</td>\n",
       "      <td>140.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>2790.0</td>\n",
       "      <td>15.6</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>394</th>\n",
       "      <td>44.0</td>\n",
       "      <td>4</td>\n",
       "      <td>97.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>2130.0</td>\n",
       "      <td>24.6</td>\n",
       "      <td>82</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>395</th>\n",
       "      <td>32.0</td>\n",
       "      <td>4</td>\n",
       "      <td>135.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>2295.0</td>\n",
       "      <td>11.6</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>396</th>\n",
       "      <td>28.0</td>\n",
       "      <td>4</td>\n",
       "      <td>120.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2625.0</td>\n",
       "      <td>18.6</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>397</th>\n",
       "      <td>31.0</td>\n",
       "      <td>4</td>\n",
       "      <td>119.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>2720.0</td>\n",
       "      <td>19.4</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      MPG  Cylinders  Displacement  Horsepower  Weight  Acceleration  \\\n",
       "393  27.0          4         140.0        86.0  2790.0          15.6   \n",
       "394  44.0          4          97.0        52.0  2130.0          24.6   \n",
       "395  32.0          4         135.0        84.0  2295.0          11.6   \n",
       "396  28.0          4         120.0        79.0  2625.0          18.6   \n",
       "397  31.0          4         119.0        82.0  2720.0          19.4   \n",
       "\n",
       "     Model Year  Origin  \n",
       "393          82       1  \n",
       "394          82       2  \n",
       "395          82       1  \n",
       "396          82       1  \n",
       "397          82       1  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = raw_dataset.copy()\n",
    "dataset.tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3MWuJTKEDM-f"
   },
   "source": [
    "### データのクレンジング\n",
    "\n",
    "このデータセットには、いくつか欠損値があります。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.572669Z",
     "iopub.status.busy": "2022-08-09T01:44:05.572090Z",
     "iopub.status.idle": "2022-08-09T01:44:05.577957Z",
     "shell.execute_reply": "2022-08-09T01:44:05.577402Z"
    },
    "id": "JEJHhN65a2VV"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MPG             0\n",
       "Cylinders       0\n",
       "Displacement    0\n",
       "Horsepower      6\n",
       "Weight          0\n",
       "Acceleration    0\n",
       "Model Year      0\n",
       "Origin          0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9UPN0KBHa_WI"
   },
   "source": [
    "この最初のチュートリアルでは簡単化のためこれらの行を削除します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.581444Z",
     "iopub.status.busy": "2022-08-09T01:44:05.580947Z",
     "iopub.status.idle": "2022-08-09T01:44:05.584938Z",
     "shell.execute_reply": "2022-08-09T01:44:05.584357Z"
    },
    "id": "4ZUDosChC1UN"
   },
   "outputs": [],
   "source": [
    "dataset = dataset.dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8XKitwaH4v8h"
   },
   "source": [
    "`\"Origin\"` 列はカテゴリであり、数値ではないので、[pd.get_dummies](https://pandas.pydata.org/docs/reference/api/pandas.get_dummies.html) でワンホットに変換します。\n",
    "\n",
    "注意: `keras.Model` を設定して、このような変換を行うことができます。これについては、このチュートリアルでは取り上げません。例については、[前処理レイヤー](../structured_data/preprocessing_layers.ipynb)または [CSV データの読み込み](../load_data/csv.ipynb)のチュートリアルをご覧ください。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.588298Z",
     "iopub.status.busy": "2022-08-09T01:44:05.587759Z",
     "iopub.status.idle": "2022-08-09T01:44:05.591980Z",
     "shell.execute_reply": "2022-08-09T01:44:05.591417Z"
    },
    "id": "gWNTD2QjBWFJ"
   },
   "outputs": [],
   "source": [
    "dataset['Origin'] = dataset['Origin'].map({1: 'USA', 2: 'Europe', 3: 'Japan'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.595418Z",
     "iopub.status.busy": "2022-08-09T01:44:05.594775Z",
     "iopub.status.idle": "2022-08-09T01:44:05.608003Z",
     "shell.execute_reply": "2022-08-09T01:44:05.607433Z"
    },
    "id": "ulXz4J7PAUzk"
   },
   "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>MPG</th>\n",
       "      <th>Cylinders</th>\n",
       "      <th>Displacement</th>\n",
       "      <th>Horsepower</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>Model Year</th>\n",
       "      <th>Europe</th>\n",
       "      <th>Japan</th>\n",
       "      <th>USA</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>393</th>\n",
       "      <td>27.0</td>\n",
       "      <td>4</td>\n",
       "      <td>140.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>2790.0</td>\n",
       "      <td>15.6</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>394</th>\n",
       "      <td>44.0</td>\n",
       "      <td>4</td>\n",
       "      <td>97.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>2130.0</td>\n",
       "      <td>24.6</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>395</th>\n",
       "      <td>32.0</td>\n",
       "      <td>4</td>\n",
       "      <td>135.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>2295.0</td>\n",
       "      <td>11.6</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>396</th>\n",
       "      <td>28.0</td>\n",
       "      <td>4</td>\n",
       "      <td>120.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2625.0</td>\n",
       "      <td>18.6</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>397</th>\n",
       "      <td>31.0</td>\n",
       "      <td>4</td>\n",
       "      <td>119.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>2720.0</td>\n",
       "      <td>19.4</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      MPG  Cylinders  Displacement  Horsepower  Weight  Acceleration  \\\n",
       "393  27.0          4         140.0        86.0  2790.0          15.6   \n",
       "394  44.0          4          97.0        52.0  2130.0          24.6   \n",
       "395  32.0          4         135.0        84.0  2295.0          11.6   \n",
       "396  28.0          4         120.0        79.0  2625.0          18.6   \n",
       "397  31.0          4         119.0        82.0  2720.0          19.4   \n",
       "\n",
       "     Model Year  Europe  Japan  USA  \n",
       "393          82       0      0    1  \n",
       "394          82       1      0    0  \n",
       "395          82       0      0    1  \n",
       "396          82       0      0    1  \n",
       "397          82       0      0    1  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.get_dummies(dataset, columns=['Origin'], prefix='', prefix_sep='')\n",
    "dataset.tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Cuym4yvk76vU"
   },
   "source": [
    "### データをトレーニング用セットとテスト用セットに分割\n",
    "\n",
    "次に、データセットをトレーニングセットとテストセットに分割します。モデルの最終評価ではテストセットを使用します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.611362Z",
     "iopub.status.busy": "2022-08-09T01:44:05.610828Z",
     "iopub.status.idle": "2022-08-09T01:44:05.615454Z",
     "shell.execute_reply": "2022-08-09T01:44:05.614851Z"
    },
    "id": "qn-IGhUE7_1H"
   },
   "outputs": [],
   "source": [
    "train_dataset = dataset.sample(frac=0.8, random_state=0)\n",
    "test_dataset = dataset.drop(train_dataset.index)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "J4ubs136WLNp"
   },
   "source": [
    "### データの観察\n",
    "\n",
    "トレーニング用セットの列のいくつかのペアの同時分布を見てみます。\n",
    "\n",
    "一番上の行を見ると、燃費 (MPG) が他のすべてのパラメータの関数であることは明らかです。他の行を見ると、それらが互いの関数であることが明らかです。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:05.618592Z",
     "iopub.status.busy": "2022-08-09T01:44:05.618384Z",
     "iopub.status.idle": "2022-08-09T01:44:07.608383Z",
     "shell.execute_reply": "2022-08-09T01:44:07.607746Z"
    },
    "id": "oRKO_x8gWKv-"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.PairGrid at 0x7f390cadddc0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 20 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.pairplot(train_dataset[['MPG', 'Cylinders', 'Displacement', 'Weight']], diag_kind='kde')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "gavKO_6DWRMP"
   },
   "source": [
    "全体の統計値も見てみましょう。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:07.612961Z",
     "iopub.status.busy": "2022-08-09T01:44:07.612421Z",
     "iopub.status.idle": "2022-08-09T01:44:07.641688Z",
     "shell.execute_reply": "2022-08-09T01:44:07.641075Z"
    },
    "id": "yi2FzC3T21jR"
   },
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MPG</th>\n",
       "      <td>314.0</td>\n",
       "      <td>23.310510</td>\n",
       "      <td>7.728652</td>\n",
       "      <td>10.0</td>\n",
       "      <td>17.00</td>\n",
       "      <td>22.0</td>\n",
       "      <td>28.95</td>\n",
       "      <td>46.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cylinders</th>\n",
       "      <td>314.0</td>\n",
       "      <td>5.477707</td>\n",
       "      <td>1.699788</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.0</td>\n",
       "      <td>8.00</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Displacement</th>\n",
       "      <td>314.0</td>\n",
       "      <td>195.318471</td>\n",
       "      <td>104.331589</td>\n",
       "      <td>68.0</td>\n",
       "      <td>105.50</td>\n",
       "      <td>151.0</td>\n",
       "      <td>265.75</td>\n",
       "      <td>455.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Horsepower</th>\n",
       "      <td>314.0</td>\n",
       "      <td>104.869427</td>\n",
       "      <td>38.096214</td>\n",
       "      <td>46.0</td>\n",
       "      <td>76.25</td>\n",
       "      <td>94.5</td>\n",
       "      <td>128.00</td>\n",
       "      <td>225.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Weight</th>\n",
       "      <td>314.0</td>\n",
       "      <td>2990.251592</td>\n",
       "      <td>843.898596</td>\n",
       "      <td>1649.0</td>\n",
       "      <td>2256.50</td>\n",
       "      <td>2822.5</td>\n",
       "      <td>3608.00</td>\n",
       "      <td>5140.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Acceleration</th>\n",
       "      <td>314.0</td>\n",
       "      <td>15.559236</td>\n",
       "      <td>2.789230</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.80</td>\n",
       "      <td>15.5</td>\n",
       "      <td>17.20</td>\n",
       "      <td>24.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Model Year</th>\n",
       "      <td>314.0</td>\n",
       "      <td>75.898089</td>\n",
       "      <td>3.675642</td>\n",
       "      <td>70.0</td>\n",
       "      <td>73.00</td>\n",
       "      <td>76.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>82.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Europe</th>\n",
       "      <td>314.0</td>\n",
       "      <td>0.178344</td>\n",
       "      <td>0.383413</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Japan</th>\n",
       "      <td>314.0</td>\n",
       "      <td>0.197452</td>\n",
       "      <td>0.398712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>USA</th>\n",
       "      <td>314.0</td>\n",
       "      <td>0.624204</td>\n",
       "      <td>0.485101</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              count         mean         std     min      25%     50%  \\\n",
       "MPG           314.0    23.310510    7.728652    10.0    17.00    22.0   \n",
       "Cylinders     314.0     5.477707    1.699788     3.0     4.00     4.0   \n",
       "Displacement  314.0   195.318471  104.331589    68.0   105.50   151.0   \n",
       "Horsepower    314.0   104.869427   38.096214    46.0    76.25    94.5   \n",
       "Weight        314.0  2990.251592  843.898596  1649.0  2256.50  2822.5   \n",
       "Acceleration  314.0    15.559236    2.789230     8.0    13.80    15.5   \n",
       "Model Year    314.0    75.898089    3.675642    70.0    73.00    76.0   \n",
       "Europe        314.0     0.178344    0.383413     0.0     0.00     0.0   \n",
       "Japan         314.0     0.197452    0.398712     0.0     0.00     0.0   \n",
       "USA           314.0     0.624204    0.485101     0.0     0.00     1.0   \n",
       "\n",
       "                  75%     max  \n",
       "MPG             28.95    46.6  \n",
       "Cylinders        8.00     8.0  \n",
       "Displacement   265.75   455.0  \n",
       "Horsepower     128.00   225.0  \n",
       "Weight        3608.00  5140.0  \n",
       "Acceleration    17.20    24.8  \n",
       "Model Year      79.00    82.0  \n",
       "Europe           0.00     1.0  \n",
       "Japan            0.00     1.0  \n",
       "USA              1.00     1.0  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.describe().transpose()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Db7Auq1yXUvh"
   },
   "source": [
    "### ラベルと特徴量の分離\n",
    "\n",
    "ラベル、すなわち目的変数を特徴量から分離します。このラベルは、モデルに予測させたい数量です。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:07.645100Z",
     "iopub.status.busy": "2022-08-09T01:44:07.644863Z",
     "iopub.status.idle": "2022-08-09T01:44:07.648882Z",
     "shell.execute_reply": "2022-08-09T01:44:07.648340Z"
    },
    "id": "t2sluJdCW7jN"
   },
   "outputs": [],
   "source": [
    "train_features = train_dataset.copy()\n",
    "test_features = test_dataset.copy()\n",
    "\n",
    "train_labels = train_features.pop('MPG')\n",
    "test_labels = test_features.pop('MPG')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "mRklxK5s388r"
   },
   "source": [
    "## 正規化\n",
    "\n",
    "統計の表を見て、それぞれの特徴量の範囲がどれほど違っているかに注目してください。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:07.652094Z",
     "iopub.status.busy": "2022-08-09T01:44:07.651609Z",
     "iopub.status.idle": "2022-08-09T01:44:07.674622Z",
     "shell.execute_reply": "2022-08-09T01:44:07.674068Z"
    },
    "id": "IcmY6lKKbkw8"
   },
   "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>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MPG</th>\n",
       "      <td>23.310510</td>\n",
       "      <td>7.728652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cylinders</th>\n",
       "      <td>5.477707</td>\n",
       "      <td>1.699788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Displacement</th>\n",
       "      <td>195.318471</td>\n",
       "      <td>104.331589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Horsepower</th>\n",
       "      <td>104.869427</td>\n",
       "      <td>38.096214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Weight</th>\n",
       "      <td>2990.251592</td>\n",
       "      <td>843.898596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Acceleration</th>\n",
       "      <td>15.559236</td>\n",
       "      <td>2.789230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Model Year</th>\n",
       "      <td>75.898089</td>\n",
       "      <td>3.675642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Europe</th>\n",
       "      <td>0.178344</td>\n",
       "      <td>0.383413</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Japan</th>\n",
       "      <td>0.197452</td>\n",
       "      <td>0.398712</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>USA</th>\n",
       "      <td>0.624204</td>\n",
       "      <td>0.485101</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     mean         std\n",
       "MPG             23.310510    7.728652\n",
       "Cylinders        5.477707    1.699788\n",
       "Displacement   195.318471  104.331589\n",
       "Horsepower     104.869427   38.096214\n",
       "Weight        2990.251592  843.898596\n",
       "Acceleration    15.559236    2.789230\n",
       "Model Year      75.898089    3.675642\n",
       "Europe           0.178344    0.383413\n",
       "Japan            0.197452    0.398712\n",
       "USA              0.624204    0.485101"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.describe().transpose()[['mean', 'std']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "-ywmerQ6dSox"
   },
   "source": [
    "スケールや値の範囲が異なる特徴量を正規化するのはよい習慣です。\n",
    "\n",
    "これが重要な理由の 1 つは、特徴にモデルの重みが掛けられるためです。したがって、出力のスケールと勾配のスケールは、入力のスケールの影響を受けます。\n",
    "\n",
    "モデルは特徴量の正規化なしで収束する可能性がありますが、正規化によりトレーニングがはるかに安定します。\n",
    "\n",
    "注意: ここでは、簡単にするため実行しますが、ワンホット特徴を正規化する利点はありません。前処理レイヤーの使用方法の詳細については、[前処理レイヤーの使用](https://www.tensorflow.org/guide/keras/preprocessing_layers)ガイドと [Keras 前処理レイヤーを使用した構造化データの分類](https://www.tensorflow.org/tutorials/structured_data/preprocessing_layers)チュートリアルを参照してください。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aFJ6ISropeoo"
   },
   "source": [
    "### 正規化レイヤー\n",
    "\n",
    "`preprocessing.Normalization` レイヤーは、その前処理をモデルに組み込むためのクリーンでシンプルな方法です。\n",
    "\n",
    "まず、レイヤーを作成します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:07.677734Z",
     "iopub.status.busy": "2022-08-09T01:44:07.677177Z",
     "iopub.status.idle": "2022-08-09T01:44:07.697680Z",
     "shell.execute_reply": "2022-08-09T01:44:07.697122Z"
    },
    "id": "JlC5ooJrgjQF"
   },
   "outputs": [],
   "source": [
    "normalizer = tf.keras.layers.Normalization(axis=-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XYA2Ap6nVOha"
   },
   "source": [
    "次にデータに `.adapt()` します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:07.701122Z",
     "iopub.status.busy": "2022-08-09T01:44:07.700648Z",
     "iopub.status.idle": "2022-08-09T01:44:11.115036Z",
     "shell.execute_reply": "2022-08-09T01:44:11.114234Z"
    },
    "id": "CrBbbjbwV91f"
   },
   "outputs": [],
   "source": [
    "normalizer.adapt(np.array(train_features))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "oZccMR5yV9YV"
   },
   "source": [
    "これにより、平均と分散が計算され、レイヤーに保存されます。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.119269Z",
     "iopub.status.busy": "2022-08-09T01:44:11.118727Z",
     "iopub.status.idle": "2022-08-09T01:44:11.123130Z",
     "shell.execute_reply": "2022-08-09T01:44:11.122453Z"
    },
    "id": "GGn-ukwxSPtx"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[   5.478  195.318  104.869 2990.252   15.559   75.898    0.178    0.197\n",
      "     0.624]]\n"
     ]
    }
   ],
   "source": [
    "print(normalizer.mean.numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "oGWKaF9GSRuN"
   },
   "source": [
    "レイヤーが呼び出されると、入力データが返され、各特徴は個別に正規化されます。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.126352Z",
     "iopub.status.busy": "2022-08-09T01:44:11.125846Z",
     "iopub.status.idle": "2022-08-09T01:44:11.138837Z",
     "shell.execute_reply": "2022-08-09T01:44:11.138149Z"
    },
    "id": "2l7zFL_XWIRu"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "First example: [[   4.    90.    75.  2125.    14.5   74.     0.     0.     1. ]]\n",
      "\n",
      "Normalized: [[-0.87 -1.01 -0.79 -1.03 -0.38 -0.52 -0.47 -0.5   0.78]]\n"
     ]
    }
   ],
   "source": [
    "first = np.array(train_features[:1])\n",
    "\n",
    "with np.printoptions(precision=2, suppress=True):\n",
    "  print('First example:', first)\n",
    "  print()\n",
    "  print('Normalized:', normalizer(first).numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6o3CrycBXA2s"
   },
   "source": [
    "## 線形回帰\n",
    "\n",
    "DNN モデルを構築する前に、単一変数および複数変数を使用した線形回帰から始めます。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lFby9n0tnHkw"
   },
   "source": [
    "### 1 つの変数\n",
    "\n",
    "単一変数の線形回帰から始めて、`Horsepower` から `MPG` を予測します。\n",
    "\n",
    "`tf.keras` を使用したモデルのトレーニングは、通常、モデルアーキテクチャを定義することから始まります。ここでは、`tf.keras.Sequential` モデルを使用します。このモデルは、[一連のステップ](https://www.tensorflow.org/guide/keras/sequential_model)を表します。\n",
    "\n",
    "単一変数の線形回帰モデルには、次の 2 つのステップがあります。\n",
    "\n",
    "- 入力 `horsepower` を正規化します。\n",
    "- 線形変換 ($y = mx+b$) を適用して、`layers.Dense` を使用して 1 つの出力を生成します。\n",
    "\n",
    "*入力*の数は、`input_shape` 引数により設定できます。また、モデルを初めて実行するときに自動的に設定することもできます。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Xp3gAFn3TPv8"
   },
   "source": [
    "まず、馬力 `Normalization` レイヤーを作成します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.142527Z",
     "iopub.status.busy": "2022-08-09T01:44:11.142048Z",
     "iopub.status.idle": "2022-08-09T01:44:11.338274Z",
     "shell.execute_reply": "2022-08-09T01:44:11.337494Z"
    },
    "id": "1gJAy0fKs1TS"
   },
   "outputs": [],
   "source": [
    "horsepower = np.array(train_features['Horsepower'])\n",
    "\n",
    "horsepower_normalizer = layers.Normalization(input_shape=[1,], axis=None)\n",
    "horsepower_normalizer.adapt(horsepower)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "4NVlHJY2TWlC"
   },
   "source": [
    "Sequential モデルを作成します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.342522Z",
     "iopub.status.busy": "2022-08-09T01:44:11.341971Z",
     "iopub.status.idle": "2022-08-09T01:44:11.383557Z",
     "shell.execute_reply": "2022-08-09T01:44:11.382866Z"
    },
    "id": "c0sXM7qLlKfZ"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Layer (type)                Output Shape              Param #   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " normalization_1 (Normalizat  (None, 1)                3         \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " ion)                                                            \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense (Dense)               (None, 1)                 2         \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total params: 5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable params: 2\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Non-trainable params: 3\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "horsepower_model = tf.keras.Sequential([\n",
    "    horsepower_normalizer,\n",
    "    layers.Dense(units=1)\n",
    "])\n",
    "\n",
    "horsepower_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eObQu9fDnXGL"
   },
   "source": [
    "このモデルは、`Horsepower` から `MPG` を予測します。\n",
    "\n",
    "トレーニングされていないモデルを最初の 10 の馬力の値で実行します。出力は良くありませんが、期待される形状が `(10,1)` であることがわかります。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.390019Z",
     "iopub.status.busy": "2022-08-09T01:44:11.389508Z",
     "iopub.status.idle": "2022-08-09T01:44:11.795764Z",
     "shell.execute_reply": "2022-08-09T01:44:11.795112Z"
    },
    "id": "UfV1HS6bns-s"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "1/1 [==============================] - ETA: 0s"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "1/1 [==============================] - 0s 361ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[ 0.524],\n",
       "       [ 0.296],\n",
       "       [-0.968],\n",
       "       [ 0.735],\n",
       "       [ 0.665],\n",
       "       [ 0.261],\n",
       "       [ 0.788],\n",
       "       [ 0.665],\n",
       "       [ 0.173],\n",
       "       [ 0.296]], dtype=float32)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "horsepower_model.predict(horsepower[:10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CSkanJlmmFBX"
   },
   "source": [
    "モデルが構築されたら、`Model.compile()` メソッドを使用してトレーニング手順を構成します。コンパイルするための最も重要な引数は、`loss` と `optimizer` です。これらは、最適化されるもの (`mean_absolute_error`) とその方法 (`optimizers.Adam` を使用)を定義するためです。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.799262Z",
     "iopub.status.busy": "2022-08-09T01:44:11.798703Z",
     "iopub.status.idle": "2022-08-09T01:44:11.810852Z",
     "shell.execute_reply": "2022-08-09T01:44:11.810226Z"
    },
    "id": "JxA_3lpOm-SK"
   },
   "outputs": [],
   "source": [
    "horsepower_model.compile(\n",
    "    optimizer=tf.optimizers.Adam(learning_rate=0.1),\n",
    "    loss='mean_absolute_error')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Z3q1I9TwnRSC"
   },
   "source": [
    "トレーニングを構成したら、`Model.fit()` を使用してトレーニングを実行します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:11.814398Z",
     "iopub.status.busy": "2022-08-09T01:44:11.813883Z",
     "iopub.status.idle": "2022-08-09T01:44:16.512109Z",
     "shell.execute_reply": "2022-08-09T01:44:16.511273Z"
    },
    "id": "-iSrNy59nRAp"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 5.74 s, sys: 1.59 s, total: 7.32 s\n",
      "Wall time: 4.69 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "history = horsepower_model.fit(\n",
    "    train_features['Horsepower'],\n",
    "    train_labels,\n",
    "    epochs=100,\n",
    "    # Suppress logging.\n",
    "    verbose=0,\n",
    "    # Calculate validation results on 20% of the training data.\n",
    "    validation_split = 0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tQm3pc0FYPQB"
   },
   "source": [
    "`history` オブジェクトに保存された数値を使ってモデルのトレーニングの様子を可視化します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.515796Z",
     "iopub.status.busy": "2022-08-09T01:44:16.515284Z",
     "iopub.status.idle": "2022-08-09T01:44:16.524189Z",
     "shell.execute_reply": "2022-08-09T01:44:16.523499Z"
    },
    "id": "YCAwD_y4AdC3"
   },
   "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>loss</th>\n",
       "      <th>val_loss</th>\n",
       "      <th>epoch</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>3.808687</td>\n",
       "      <td>4.209890</td>\n",
       "      <td>95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>3.805177</td>\n",
       "      <td>4.200547</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>3.801378</td>\n",
       "      <td>4.182535</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>3.803104</td>\n",
       "      <td>4.174462</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>3.803327</td>\n",
       "      <td>4.186305</td>\n",
       "      <td>99</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        loss  val_loss  epoch\n",
       "95  3.808687  4.209890     95\n",
       "96  3.805177  4.200547     96\n",
       "97  3.801378  4.182535     97\n",
       "98  3.803104  4.174462     98\n",
       "99  3.803327  4.186305     99"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hist = pd.DataFrame(history.history)\n",
    "hist['epoch'] = history.epoch\n",
    "hist.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.527667Z",
     "iopub.status.busy": "2022-08-09T01:44:16.527105Z",
     "iopub.status.idle": "2022-08-09T01:44:16.531229Z",
     "shell.execute_reply": "2022-08-09T01:44:16.530572Z"
    },
    "id": "9E54UoZunqhc"
   },
   "outputs": [],
   "source": [
    "def plot_loss(history):\n",
    "  plt.plot(history.history['loss'], label='loss')\n",
    "  plt.plot(history.history['val_loss'], label='val_loss')\n",
    "  plt.ylim([0, 10])\n",
    "  plt.xlabel('Epoch')\n",
    "  plt.ylabel('Error [MPG]')\n",
    "  plt.legend()\n",
    "  plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.534365Z",
     "iopub.status.busy": "2022-08-09T01:44:16.533758Z",
     "iopub.status.idle": "2022-08-09T01:44:16.647211Z",
     "shell.execute_reply": "2022-08-09T01:44:16.646429Z"
    },
    "id": "yYsQYrIZyqjz"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CMNrt8X2ebXd"
   },
   "source": [
    "後で使用するために、テスト用セットの結果を収集します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.651004Z",
     "iopub.status.busy": "2022-08-09T01:44:16.650401Z",
     "iopub.status.idle": "2022-08-09T01:44:16.721307Z",
     "shell.execute_reply": "2022-08-09T01:44:16.720515Z"
    },
    "id": "kDZ8EvNYrDtx"
   },
   "outputs": [],
   "source": [
    "test_results = {}\n",
    "\n",
    "test_results['horsepower_model'] = horsepower_model.evaluate(\n",
    "    test_features['Horsepower'],\n",
    "    test_labels, verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "F0qutYAKwoda"
   },
   "source": [
    "これは単一変数の回帰であるため、入力の関数としてモデルの予測を簡単に確認できます。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.725816Z",
     "iopub.status.busy": "2022-08-09T01:44:16.725139Z",
     "iopub.status.idle": "2022-08-09T01:44:16.879391Z",
     "shell.execute_reply": "2022-08-09T01:44:16.878428Z"
    },
    "id": "xDS2JEtOn9Jn"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "1/8 [==>...........................] - ETA: 0s"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "8/8 [==============================] - 0s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "x = tf.linspace(0.0, 250, 251)\n",
    "y = horsepower_model.predict(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.883475Z",
     "iopub.status.busy": "2022-08-09T01:44:16.882883Z",
     "iopub.status.idle": "2022-08-09T01:44:16.887517Z",
     "shell.execute_reply": "2022-08-09T01:44:16.886869Z"
    },
    "id": "rttFCTU8czsI"
   },
   "outputs": [],
   "source": [
    "def plot_horsepower(x, y):\n",
    "  plt.scatter(train_features['Horsepower'], train_labels, label='Data')\n",
    "  plt.plot(x, y, color='k', label='Predictions')\n",
    "  plt.xlabel('Horsepower')\n",
    "  plt.ylabel('MPG')\n",
    "  plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:16.890719Z",
     "iopub.status.busy": "2022-08-09T01:44:16.890189Z",
     "iopub.status.idle": "2022-08-09T01:44:17.032403Z",
     "shell.execute_reply": "2022-08-09T01:44:17.031686Z"
    },
    "id": "7l9ZiAOEUNBL"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_horsepower(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Yk2RmlqPoM9u"
   },
   "source": [
    "### 複数の入力"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "PribnwDHUksC"
   },
   "source": [
    "ほぼ同じ設定を使用して、複数の入力に基づく予測を実行することができます。このモデルでは、$m$ が行列で、$b$ がベクトルですが、同じ $y = mx+b$ を実行します。\n",
    "\n",
    "ここでは、データセット全体に適合した `Normalization` レイヤーを使用します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:17.036454Z",
     "iopub.status.busy": "2022-08-09T01:44:17.035855Z",
     "iopub.status.idle": "2022-08-09T01:44:17.059037Z",
     "shell.execute_reply": "2022-08-09T01:44:17.058397Z"
    },
    "id": "ssnVcKg7oMe6"
   },
   "outputs": [],
   "source": [
    "linear_model = tf.keras.Sequential([\n",
    "    normalizer,\n",
    "    layers.Dense(units=1)\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IHlx6WeIWyAr"
   },
   "source": [
    "入力のバッチでこのモデルを呼び出すと、各例に対して `units=1` 出力が生成されます。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:17.062523Z",
     "iopub.status.busy": "2022-08-09T01:44:17.062030Z",
     "iopub.status.idle": "2022-08-09T01:44:17.157538Z",
     "shell.execute_reply": "2022-08-09T01:44:17.156788Z"
    },
    "id": "DynfJV18WiuT"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "1/1 [==============================] - ETA: 0s"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "1/1 [==============================] - 0s 50ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[ 0.34 ],\n",
       "       [-0.635],\n",
       "       [ 1.53 ],\n",
       "       [-1.483],\n",
       "       [-0.969],\n",
       "       [-0.55 ],\n",
       "       [-1.092],\n",
       "       [-1.881],\n",
       "       [ 0.499],\n",
       "       [ 0.644]], dtype=float32)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "linear_model.predict(train_features[:10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hvHKH3rPXHmq"
   },
   "source": [
    "モデルを呼び出すと、その重み行列が作成されます。これで、`kernel` ($y=mx+b$ の $m$) の形状が (9,1) であることがわかります。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:17.161014Z",
     "iopub.status.busy": "2022-08-09T01:44:17.160422Z",
     "iopub.status.idle": "2022-08-09T01:44:17.167903Z",
     "shell.execute_reply": "2022-08-09T01:44:17.167340Z"
    },
    "id": "DwJ4Fq0RXBQf"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Variable 'dense_1/kernel:0' shape=(9, 1) dtype=float32, numpy=\n",
       "array([[ 0.264],\n",
       "       [-0.682],\n",
       "       [ 0.315],\n",
       "       [ 0.405],\n",
       "       [-0.112],\n",
       "       [-0.742],\n",
       "       [-0.734],\n",
       "       [-0.378],\n",
       "       [-0.531]], dtype=float32)>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "linear_model.layers[1].kernel"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eINAc6rZXzOt"
   },
   "source": [
    "Keras `Model.compile` でモデルを構成し、`Model.fit` で 100 エポックトレーニングします。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:17.170945Z",
     "iopub.status.busy": "2022-08-09T01:44:17.170705Z",
     "iopub.status.idle": "2022-08-09T01:44:17.178324Z",
     "shell.execute_reply": "2022-08-09T01:44:17.177689Z"
    },
    "id": "A0Sv_Ybr0szp"
   },
   "outputs": [],
   "source": [
    "linear_model.compile(\n",
    "    optimizer=tf.optimizers.Adam(learning_rate=0.1),\n",
    "    loss='mean_absolute_error')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:17.181472Z",
     "iopub.status.busy": "2022-08-09T01:44:17.181187Z",
     "iopub.status.idle": "2022-08-09T01:44:21.919694Z",
     "shell.execute_reply": "2022-08-09T01:44:21.918989Z"
    },
    "id": "EZoOYORvoTSe"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 5.91 s, sys: 1.53 s, total: 7.44 s\n",
      "Wall time: 4.73 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "history = linear_model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    epochs=100,\n",
    "    # Suppress logging.\n",
    "    verbose=0,\n",
    "    # Calculate validation results on 20% of the training data.\n",
    "    validation_split = 0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EdxiCbiNYK2F"
   },
   "source": [
    "この回帰モデルですべての入力を使用すると、入力が 1 つだけの `horsepower_model` よりもトレーニングエラーや検証エラーが大幅に低くなります。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:21.923272Z",
     "iopub.status.busy": "2022-08-09T01:44:21.922991Z",
     "iopub.status.idle": "2022-08-09T01:44:22.047138Z",
     "shell.execute_reply": "2022-08-09T01:44:22.046414Z"
    },
    "id": "4sWO3W0koYgu"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjIAAAG2CAYAAABlBWwKAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAABUYElEQVR4nO3deXxU9b3/8deZJZN9JxsECBB2RAREwF2s4obW2p+KLdbeWisulNYqte61qG2prQteeyu9vVXx2or1uhZRRJR93wmL7Nn3fTJzfn9MMhhJwgQyW/J+Ph55TObMOZNPvgTy5nu+i2GapomIiIhIGLIEuwARERGRU6UgIyIiImFLQUZERETCloKMiIiIhC0FGREREQlbCjIiIiISthRkREREJGwpyIiIiEjYUpARERGRsKUgIyIiImErqEFm2bJlXH311WRlZWEYBm+//Xar103T5OGHHyYzM5OoqCimTJlCXl5ecIoVERGRkBPUIFNTU8Po0aN54YUX2nz9mWee4U9/+hMvvfQSq1atIiYmhssuu4z6+voAVyoiIiKhyAiVTSMNw2DRokVce+21gKc3Jisri5/97Gf8/Oc/B6CiooL09HT++te/cuONNwaxWhEREQkFtmAX0J79+/eTn5/PlClTvMcSEhKYMGECK1asaDfINDQ00NDQ4H3udrspLS0lJSUFwzD8XreIiIicPtM0qaqqIisrC4ul/RtIIRtk8vPzAUhPT291PD093ftaW+bOnctjjz3m19pEREQkMA4dOkSfPn3afT1kg8ypmjNnDrNnz/Y+r6iooG/fvuzfv5+4uLgu+zpOp5NPP/2Uiy66CLvd3mXv2118srOQn/1jK8Mz43j1h+M7PNey8gWsy3+He9i1uK78Q5vnqL0DR20dOGrrwFFbB05XtXVVVRU5OTkn/d0dskEmIyMDgIKCAjIzM73HCwoKOPPMM9u9zuFw4HA4TjienJxMfHx8l9XndDqJjo4mJSVFfynaMCY3AotjH4eqPW3f4W29geNgjQG1+yAlpc1T1N6Bo7YOHLV14KitA6er2rrl2pMNCwnZdWRycnLIyMhgyZIl3mOVlZWsWrWKiRMnBrEy8UXf5BisFoOaRhcFlQ0dn5w2zPNYtBtcTf4vTkREuo2g9shUV1ezZ88e7/P9+/ezceNGkpOT6du3L7NmzeLXv/41ubm55OTk8NBDD5GVleWd2SShK8JmoW9yNPuLa9hbVE1GQmT7Jyf2A3s0OGuhbD+k5gauUBERCWtB7ZFZu3YtY8aMYcyYMQDMnj2bMWPG8PDDDwPwi1/8grvvvpvbb7+d8ePHU11dzYcffkhkZAe/FCVkDOwVA8CewuqOT7RYoNdQz+eF2/1clYiIdCdB7ZG58MIL6WgZG8MwePzxx3n88ccDWJV0lUFpcXy8o/DkQQYgbTgcXQ+FO2D4NP8XJyI9itvtxmazUV9fj8vlCnY53ZrT6fSpre12O1ar9bS/XsgO9pXwNygtFoC8wqqTn9wyTkY9MiLSxRobG9m/fz8ZGRkcOnRIa4r5mWmaPrd1YmIiGRkZp/VnoiAjfpPbHGT2FNac/GRvkNnhx4pEpKcxTZNjx45htVrp06cPcXFxHS6uJqfP7XZTXV1NbGxsu21tmia1tbUUFhYCtJqd3FkKMuI3A5uDTHF1A+W1jSRGR7R/ctpwz2PJXnDWg13joETk9DU1NVFbW+v9RRkZGakg42dut5vGxsaTtnVUVBQAhYWFpKWlnfJtJv1pit/EOmxkNc9WOuk4mbgMiEoC0wXFuwNQnYj0BC1jNLR2TGiKjo4GPONqTpWCjPjVoHTPiox5JwsyhnG8V0a3l0Ski2lcTGjqij8XBRnxq5ZxMnkFvsxc0oBfERHpHAUZ8auWmUt7ijoTZNQjIyI924UXXsisWbOCXUZYUJARv/LOXCrwZQq2bi2JiEjnKMiIX7X0yBytqKe64ST7KLWs7ltxEOor/VyZiIh0Bwoy4leJ0RGkxnp2I997sgG/0ckQ17yWQNEuP1cmIhIeysrK+P73v09SUhLR0dFMnTqVvLw87+sHDhzg6quvJikpiZiYGEaMGMH777/vvXb69On06tWLqKgocnNzWbBgQbC+Fb/QOjLid7lpsRRXN5BXWM3o7MSOT04bBlXHPAN+s8cHpD4R6TlM06S28SS9w34SZbee0iydW2+9lby8PN555x3i4+O5//77ueKKK9i+fTt2u52ZM2fS2NjIsmXLiImJYfv27cTGenrDH3roIbZv384HH3xAamoqe/bsoa6urqu/taBSkBG/G5QWy4p9JT5uVTAc9n6icTIi4hd1ThcjH10clK+9/fHLiI7o3K/dlgDzxRdfMGnSJABeffVVsrOzefvtt7nhhhs4ePAg119/PaNGjQJgwIAB3usPHjzImDFjGDduHAD9+/fvmm8mhOjWkvhdbrrnfwYnvbUEX5u5tM2PFYmIhIcdO3Zgs9mYMGGC91hKSgpDhgxhxw7Pf/juuecefv3rXzN58mQeeeQRNm/e7D33Jz/5CQsXLuTMM8/kF7/4BV9++WXAvwd/U4+M+N3xzSN93AUb1CMjIn4RZbey/fHLgva1/eE//uM/uOyyy3jvvff497//zdy5c/n973/P3XffzdSpUzlw4ADvv/8+ixcv5pJLLmHmzJn87ne/80stwaAeGfG7liBzqLSWemf7W7oD0GsIYEBNEVQX+r84EelRDMMgOsIWlI9TGR8zbNgwmpqaWLVqlfdYSUkJu3btYvjw4d5j2dnZ3HHHHbz11lv87Gc/489//rP3tV69ejFjxgz+/ve/8+yzz/Lyyy+fXiOGGAUZ8btesQ4Souy4TdhXdJKdsCNiILn5/m7+Fv8XJyISwnJzc5k2bRo/+tGPWL58OZs2beKWW26hd+/eTJs2DYBZs2bx0UcfsX//ftavX8+nn37KsGGe2/QPP/ww//rXv9izZw/btm3j3Xff9b7WXSjIiN8ZhvG120s+DPjN8AxYo2CrH6sSEQkPCxYsYOzYsVx11VVMnDgR0zR5//33vRthulwuZs6cybBhw7j88ssZPHgwL774IgARERHMmTOHM844g/PPPx+r1crChQuD+e10OY2RkYDITYtl3YEy3wb8ZoyE7W9DvoKMiPRMS5cu9X6elJTE3/72t3bPfe6559p97Ve/+hW/+tWvurK0kKMeGQmITg34TVePjIiI+EZBRgKiU0EmY6TnsXg3NDX4sSoREQl3CjISELnpcQB8VVyD0+Xu+OT43hCZCO4mKNrp/+JERCRsKchIQGQlRBITYaXJbXKg5CQzlwzj+IBfjZMREZEOKMhIQBiGwcCW20sFvoyTab69pHEyIiLSAQUZCZiWcTJ7OjNORmvJiIhIBxRkJGA6N3Ppaz0ypunHqkREJJwpyEjA5KZ5Bvz61CPTaygYVqgrg8ojfq5MRETClYKMBExuc4/M3qJqXO6T9LLYIyF1sOdzDfgVEZF2KMhIwGQnRxNhs9DQ5OZIWd3JL2gZJ1OgcTIiIp3Rv39/nn32WZ/ONQyDt99+26/1+JOCjASM1WIwIDUG8HHPpZZxMuqRERGRdijISECd0swlTcEWEZF2KMhIQHUqyLTsuVSyFxpPsoieiEg38fLLL5OVlYXb3XoV9GnTpnHbbbexd+9epk2bRnp6OrGxsYwfP56PP/64y77+li1buPjii4mKiiIlJYXbb7+d6urj/2YvXbqUs88+m5iYGBITE5k8eTIHDhwAYNOmTVxyySVkZ2eTmJjI2LFjWbt2bZfV1hYFGQmoTk3BjkuHmDTAxNBWBSLSFUzT8x+jYHz4uJTEDTfcQElJCZ9++qn3WGlpKR9++CHTp0+nurqaK664giVLlrBhwwYuv/xyrr76ag4ePHjazVNTU8Nll11GUlISa9as4c033+Tjjz/mrrvuAqCpqYlrr72WCy64gM2bN7NixQpuv/12DMMAYPr06fTu3ZslS5awZs0aHnjgAex2+2nX1RGbX99d5BtapmDvLazGNE3vD3+7MkbC3k+aby/18n+BItK9OWvhqT7B+dq/PAoRMSc9LSkpialTp/Laa69xySWXAPCPf/yD1NRULrroIiwWC6NHj/ae/8QTT7Bo0SLeeecdb+A4Va+99hr19fX87W9/IybGU+vzzz/P1VdfzdNPP43dbqeiooKrrrqKgQMHAjBs2DDv9QcPHuRnP/sZgwcPJj4+niFDhpxWPb5Qj4wEVP/UaCwGVDU0UVjlw87WzQN+jcJtfq5MRCR0TJ8+nX/+8580NHj+nXz11Ve58cYbsVgsVFdX8/Of/5xhw4aRmJhIbGwsO3bs6JIemR07djB69GhviAGYPHkybrebXbt2kZyczK233spll13G1VdfzR//+EeOHTvmPXf27NncfvvtXHvttTz99NPs3bv3tGs6GfXISEA5bFb6pcSwv7iGvIJq0uMjO76gefNIo2Ab9LrQ/wWKSPdmj/b0jATra/vo6quvxjRN3nvvPcaPH8/nn3/OH/7wBwB+/vOfs3jxYn73u98xaNAgoqKi+M53vkNjY6O/Km9lwYIF3HPPPXz44Ye88cYb/OpXv2Lx4sWcc845PProo9x444289dZbfPLJJzz66KMsXLiQ6667zm/1qEdGAu74gF/fp2AbhdvAdJ/kZBGRkzAMz+2dYHyc7Fb610RGRvLtb3+bV199lddff50hQ4Zw1llnAfDFF19w6623ct111zFq1CgyMjL46quvuqR5hg0bxqZNm6ipOT7B4osvvsBisbS6TTRmzBjmzJnDl19+yciRI3nttde8rw0ePJg777yTjz76iG9/+9ssWLCgS2prj4KMBJw3yBT5MOA3NResERiN1UQ3Fvm5MhGR0DF9+nTee+89XnnlFaZPn+49npuby1tvvcXGjRvZtGkTN9988wkznE7na0ZGRjJjxgy2bt3Kp59+yt133833vvc90tPT2b9/P3PmzGHFihUcOHCAf//73+Tl5TFs2DDq6uq46667WLp0KQcPHuSLL75gzZo1rcbQ+INuLUnA5XZmCrbV7tl3KX8zCXWnf/9XRCRcXHzxxSQnJ7Nr1y5uvvlm7/F58+Zx2223MWnSJFJTU7n//vuprKzskq8ZHR3NRx99xL333sv48eOJjo7m+uuvZ968ed7Xd+7cyX//939TUlJCZmYmM2fO5Mc//jFNTU2UlJRw6623UlBQQGpqKt/+9rd57LHHuqS29ijISMB1ai0Z8IyTyd9MfN0hP1YlIhJaLBYLR4+eOJ6nf//+fPLJJ62OzZw5s9XzztxqMr8xLXzUqFEnvH+L9PR0Fi1a1OZrERERvP7667jdbiorK4mPj8di8f+NH91akoAb2MsTZIqrGymv9WFwWvOAX/XIiIjINynISMDFOGz0TowCfF3h1zPgV0FGRKRzXn31VWJjY9v8GDFiRLDL6xK6tSRBMTAtliPldeQVVjOuf3LHJzfvuRTdWIyzvgLsqQGoUEQk/F1zzTVMmDChzdf8veJuoCjISFAM6hXLst1FvvXIRCVhJmRjVBzCKNgKcRf6vT4Rke4gLi6OuLi4YJfhV7q1JEGRm965Ab9my3oy2glbRE7BNwe0Smjoij8XBRkJis7OXDLTPPdyFWREpDOsVisATqczyJVIW2pra4HTu82lW0sSFIOaZy4dKa+jpqGJGEfHP4pmestWBQoyIuI7m81GdHQ0RUVFxMfHU19fH5ApwT2Z2+2msbGxw7Y2TZPa2loKCwtJTEz0Bs5ToSAjQZEUE0FqbATF1Y3sK6phVJ+EDs83mwf8UrwLmhrBFhGAKkUk3BmGQWZmJvv27ePw4cNERUVhdGKrAOk80zSpq6vzqa0TExPJyMg4ra+nICNBM7BXLMXVpeQVVp00yJDQF6clCrurDop3e2cyiYicTEREBDk5OSxevJgLLrig28zWCVVOp5Nly5Zx/vnnd9jWdrv9tHpiWijISNDkpseyan+pb+NkDIOK6L6kVu+C/C0KMiLSKRaLBZfLRWRkpIKMn1mtVpqamgLW1rpRKEHTMk7G1wG/lVF9PZ9onIyIiDRTkJGgGZTmWdvAp12wgYqWIJO/2V8liYhImFGQkaBpWUvmQEktjU0n34K+Iqqf55P8LaA1IUREBAUZCaK0OAdxDhsut8lXJTUnPb8qMgvTYoO6Mqg8EoAKRUQk1CnISNAYhsHATiyM57ZEQEqu50m+xsmIiIiCjARZbnOQ2V1Q5dP5LVsVkL/FXyWJiEgYUZCRoBqS4Rnwuyu/s0FGA35FRERBRoKs80HGs1WBemRERAQUZCTIWoLMVyU11DtdJz3fTPdsHknZfmjwLfyIiEj3pSAjQdUr1kFyTARuE/IKfFhPJjoF4rI8nxds829xIiIS8hRkJKgMw2BIuqdXZmd+pW8XZej2koiIeCjISNB1dpzM8SCjAb8iIj2dgowE3dCWIOPjFOzjQUZryYiI9HQKMhJ0LT0yOzvbI1O4HVxNfqpKRETCQUgHGZfLxUMPPUROTg5RUVEMHDiQJ554AlP77HQrg5vHyBRVNVBa03jyC5JyICIWmuqhJM/P1YmISCgL6SDz9NNPM3/+fJ5//nl27NjB008/zTPPPMNzzz0X7NKkC8U4bPRNjgZ8HPBrsUDLwnjHNE5GRKQnC+kg8+WXXzJt2jSuvPJK+vfvz3e+8x2+9a1vsXr16mCXJl2s0wN+M0d7Ho9t8lNFIiISDmzBLqAjkyZN4uWXX2b37t0MHjyYTZs2sXz5cubNm9fuNQ0NDTQ0NHifV1Z6/ofvdDpxOp1dVlvLe3Xle/Zkub1iWAzsOFrRZpt+s72NtBHYAPexjbj0Z9Cl9LMdOGrrwFFbB05XtbWv1xtmCA84cbvd/PKXv+SZZ57BarXicrl48sknmTNnTrvXPProozz22GMnHH/ttdeIjo72Z7lyGjYUG/w1z0q/WJPZo06+wm987UEu2vUrnNZo3h81HwwjAFWKiEig1NbWcvPNN1NRUUF8fHy754V0kFm4cCH33Xcfv/3tbxkxYgQbN25k1qxZzJs3jxkzZrR5TVs9MtnZ2RQXF3fYEJ3ldDpZvHgxl156KXa7vcvet6faU1jN1Oe+JDrCyoYHL8ZiaR1MTmhvlxPbb/thuBpx3rkWkvoHp/BuSD/bgaO2Dhy1deB0VVtXVlaSmpp60iAT0reW7rvvPh544AFuvPFGAEaNGsWBAweYO3duu0HG4XDgcDhOOG632/3yw+uv9+1pcjMSiLBZqG10UVDdRN+UtnvPvO1tt0PacDi2EXvRNkjLDXDF3Z9+tgNHbR04auvAOd229vXakB7sW1tbi8XSukSr1Yrb7Q5SReIvNquFQb1igU5sVZB5hudRK/yKiPRYIR1krr76ap588knee+89vvrqKxYtWsS8efO47rrrgl2a+MFQzVwSEZFOCulbS8899xwPPfQQd955J4WFhWRlZfHjH/+Yhx9+ONiliR94V/j1eauCrwUZ09SAXxGRHiikg0xcXBzPPvsszz77bLBLkQDo9Foy6SPAsEBNEVTlQ3ymH6sTEZFQFNK3lqRnGZrhGZW+v7iGhqaTT8EmIhpSB3s+1zgZEZEeSUFGQkZ6vIOEKDsut8mewmrfLtI4GRGRHk1BRkKGYRidv72U0TxzSUFGRKRHUpCRkHLqM5d0a0lEpCdSkJGQ0tIjs8PnHplRnseKg1Bb6qeqREQkVCnISEg53iPj46J4UYnHtyfQgF8RkR5HQUZCyuB0T5ApqGygtKbRt4s0TkZEpMdSkJGQEhdpp1/zPkvbj/q6VYHGyYiI9FQKMhJyRmYlALD1aIVvF2gKtohIj6UgIyFnRG/Pwnhbj3QyyJTsgQYf158REZFuQUFGQs6I5h4Zn28txaZBbAZgQsFW/xUmIiIhR0FGQs6ILE+PzL7iGqrqnb5dpHEyIiI9koKMhJzUWAeZCZEA7Djm68J4mrkkItITKchISGrpldnm64DflinYBVv8VJGIiIQiBRkJSS3jZLYe8XGcTPoIz2PhTnA1+akqEREJNQoyEpI63SOTlAP2GHA1eGYviYhIj6AgIyFpZG9Pj0xeYTX1TtfJL7BYIH2453PNXBIR6TEUZCQkZSZEkhwTgctt+r4TdvpIz6OCjIhIj6EgIyHJMIyv3V7ycZxMRnOQyVeQERHpKRRkJGSN6OxWBd4emW1+qkhEREKNgoyErJHNWxVs83WrgrTmMTJVR6G21E9ViYhIKFGQkZDV0iOzI78Kp8t98gsi4yGxn+dzjZMREekRFGQkZPVLjibWYaOxyc2+ohrfLsoY5XnUOBkRkR5BQUZClsViMLxlwO+xTi6Mp3EyIiI9goKMhLSWmUvbfd1zyTvgV1sViIj0BAoyEtJGNo+T8XkKtrYqEBHpURRkJKS1rPC741gVbtOHC7RVgYhIj6IgIyFtYK8YHDYLNY0uiut9uEBbFYiI9CgKMhLSbFYLQzM942QO1xi+XaStCkREegwFGQl5I5sH/B7yOcho5pKISE+hICMhb1TzOJlD1T5eoLVkRER6DAUZCXmjsxMBOFhj4PZlxK+2KhAR6TEUZCTk5abFEmm30OAy2Ffswwq/2qpARKTHUJCRkGezWrzjZLYc8XU9Ge2ELSLSEyjISFg4o3mczGZfd8LOaA4yGicjItKtKchIWBjV29Mjs/mwj0HGO3NJQUZEpDtTkJGwcEaf5hV+86toaHKd/IKWW0uFO7RVgYhIN6YgI2EhOymKGJuJ02Wy05cNJLVVgYhIj6AgI2HBMAz6xnqmXm86XH7yCyyW4+vJHN3gv8JERCSoFGQkbPSN9TxuOuTjOJk+4zyPh1f7pyAREQk6BRkJG53qkQHIPtvzeGiNfwoSEZGgU5CRsNE3xhNk9hZVU1XvPPkFfZqDTOE2aPBhXI2IiIQdBRkJG/ER0DsxEtOELb6sJxOfCQnZYLrhyHr/FygiIgGnICNhpWUDSd/HyYz3PGqcjIhIt6QgI2HljD4tC+OV+3aBxsmIiHRrCjISVs7w9siU+3ZByziZw2vA9GHnbBERCSsKMhJWRmTFYxhwtKKewqr6k1+QMQpskVBXCiV7/V+giIgElIKMhJVYh43cNM+CMpt9GSdji4DMMz2fa5yMiEi3oyAjYWd0n0SgM+vJNA/4PaQgIyLS3SjISNg5IzsRgE2+7oT99XEyIiLSrSjISNg5s6VH5lA5pi8DeFtmLhVu18J4IiLdjIKMhJ0hGXFE2CxU1Dn5qqT25BfEZUBC3+aF8db5v0AREQkYBRkJOxE2CyOzPOvJbDhY5ttF3nEyur0kItKdKMhIWDqrbxIA630NMt5xMhrwKyLSnSjISFg6q19zkDlQ7tsFLT0yWhhPRKRbUZCRsNTSI7Mzv5KahqaTX5DesjBeGZTs8XN1IiISKAoyEpYyEiLJSojEbfq4nszXF8bTejIiIt2GgoyErTHNt5c2HCz37YJs7YQtItLdKMhI2PIO+D3Q2QG/moItItJdKMhI2DqrbyIAG3xdGC9ztOexaCe4nP4rTEREAkZBRsLWiKwEImwWSmsaOeDLwniJfSEiDtxOKM7zf4EiIuJ3CjIStiJsFkb1TgB8XE/GMCB9hOfzgq1+rExERAJFQUbC2pjmDSR9XhgvY6TnUUFGRKRbCPkgc+TIEW655RZSUlKIiopi1KhRrF27NthlSYjo9MJ43h6Zbf4pSEREAsoW7AI6UlZWxuTJk7nooov44IMP6NWrF3l5eSQlJQW7NAkR31wYL8Zxkh/p9OYemXz1yIiIdAchHWSefvppsrOzWbBggfdYTk5OECuSUNOyMN7Rino2HS5n0sDUji9IG+55rM6HmmKIOcn5IiIS0nwKMu+8806n3/jSSy8lKiqq09d98+tedtll3HDDDXz22Wf07t2bO++8kx/96EftXtPQ0EBDQ4P3eWVlJQBOpxOns+um3La8V1e+p7Svo/Y+MzuBoxX1rN1fwvi+CR2/kcWBLSkHo2w/TUc3YfY/3x/lhjX9bAeO2jpw1NaB01Vt7ev1hunDAhwWS+eG0hiGQV5eHgMGDOjUdd8UGRkJwOzZs7nhhhtYs2YN9957Ly+99BIzZsxo85pHH32Uxx577ITjr732GtHR0adVj4SmpccMFn1lZUSSm9uHuk96/vh9fySrYh1bet/MvrTLA1ChiIh0Vm1tLTfffDMVFRXEx8e3e57PQSY/P5+0tDSfvnhcXBybNm067SATERHBuHHj+PLLL73H7rnnHtasWcOKFSvavKatHpns7GyKi4s7bIjOcjqdLF68mEsvvRS73d5l7ytt66i9Nx4q54aXV5MUbWfVAxdiGEaH72VZ9gzWz5/BfcZNuK5+zp9lhyX9bAeO2jpw1NaB01VtXVlZSWpq6kmDjE+3lmbMmNGp20S33HJLl4SGzMxMhg8f3urYsGHD+Oc//9nuNQ6HA4fDccJxu93ulx9ef72vtK2t9h7dN4UIm4WyWidHKp3kpMZ0/CZZZwBgKdqORX927dLPduCorQNHbR04p9vWvl7r0z2jBQsWEBcX5/MXnz9/Pqmppz+IcvLkyezatavVsd27d9OvX7/Tfm/pPlotjOfLvkstU7ALd4KryY+ViYiIv4X0OjI//elPWblyJb/5zW/Ys2cPr732Gi+//DIzZ84MdmkSYlr2XfJpYbzE/hARC64GKNnj17pERMS/fA4yx44d48EHH/Q+P/fccznrrLO8H+PHj+fIkSNdWtz48eNZtGgRr7/+OiNHjuSJJ57g2WefZfr06V36dST8jeufDMCKfSUnP9liOT4NWyv8ioiENZ+DzIsvvkhZ2fH/7W7atInzzjuPadOmMW3aNKxWK3/4wx+6vMCrrrqKLVu2UF9fz44dOzqcei0918SBKVgtBvuKajhc5sMGktpzSUSkW/B5Qbx3332XP/3pT62O3Xvvvd6ZSeeccw6zZ8/md7/7XddWKOKD+Eg7Y7ITWXugjM/zirnp7L4dX+Ddc0lbFYiIhDOfe2S++uqrVqvqXnrppcTEHJ8dMmTIEPbv39+11Yl0wvmDewGwbHfRyU9OV5AREekOfA4yTqeToqLjvyDeeust0tPTvc/Lyso6vXCeSFc6L9czU+6LPcU0uU6yMF7LGJnKI1Bb6ufKRETEX3xOHkOGDGm1MN03ff755wwePLhLihI5FWf0SSQhyk5lfRObj1R0fHJkPCQ2T+NXr4yISNjyOcjceOONPPzww2zevPmE1zZt2sTjjz/OTTfd1KXFiXSG1WJw7iBPr4xuL4mI9Aw+D/adNWsW7777LmPHjuXSSy9lyJAhAOzatYvFixczceJEZs2a5a86RXxyXm4q7205xud5xcyacpIewvQRsOs9KNgSmOJERKTL+Rxk7HY7ixcvZt68eSxcuJClS5cCkJubyxNPPMFPf/pTLfssQXde84DfjYfKqahzkhDVwc+kZi6JiIQ9n4MMeDZxfOCBB3jggQf8VY/IaemdGMXAXjHsLaphxd5iLh+Z2f7JLbeWCneA2wUWa2CKFBGRLtOpaUZvvPEG06dP54YbbuCll17yV00ip+W83OZp2HnFHZ+Y1B/s0dBUDyV7/V+YiIh0OZ+DzPz587nppptYu3YteXl53Hnnndx3333+rE3klFzwtfVkTNNs/0SL9WtbFWicjIhIOPI5yDz//PM88sgj7Nq1i40bN/K3v/2NF1980Z+1iZySCQOSibBaOFxWx1clJ9muIL0lyGz3f2EiItLlfA4y+/btY8aMGd7nN998M01NTRw7dswvhYmcqugIG+P6JwE+TMP2jpNRkBERCUc+B5mGhoZWWxJYLBYiIiKoq6vzS2Eip6NlnMzneScLMto8UkQknHVq1tJDDz1EdHS093ljYyNPPvkkCQkJ3mPz5s3ruupETtF5uak8/SGs2FtCY5ObCFs7mb1ljEz5Qaiv9Kz4KyIiYcPnIHP++eeza9euVscmTZrEvn37vM8Nw+i6ykROw/DMeFJjIyiubmTDwTImDEhp+8ToZIjLgqqjnmnYfScEtlARETktPgeZlgXwRMKBxWIwaWAq72w6ypd7S9oPMuC5vVR11HN7SUFGRCSsaLtq6bYmDfSElxV7Szo+sWXmkgb8ioiEHZ97ZB5//HGfznv44YdPuRiRrjRpoGcDyQ2HyqhtbCI6op0fd20eKSIStnwOMo8++ihZWVmkpaW1u8iYYRgKMhIyspOj6J0YxZHyOtZ+Vcb5zQvlnSDta2vJmCZorJeISNjwOchMnTqVTz75hHHjxnHbbbdx1VVXYbHozpSELsMwmDQwhTfXHebLvSXtB5nUwWCxQUMFVByGxOzAFioiIqfM5yTy3nvvsXfvXiZMmMB9991H7969uf/++0+YySQSSiZ6x8l0sO+SLQJSh3g+1+0lEZGw0qkulaysLObMmcOuXbt44403KCwsZPz48UyePFkL40lIagkyW45UUFHnbP9E74BfBRkRkXByyveGxo8fz0UXXcSwYcPYsGEDTmcHvyREgiQzIYoBqTG4TVi9v7T9E70r/CrIiIiEk04HmRUrVvCjH/2IjIwMnnvuOWbMmMHRo0eJj9eKqBKaWnplvuzo9lJaS5DRFGwRkXDic5B55plnGD58ONOmTSM2NpbPP/+cNWvWcOedd5KYmOjHEkVOT8s07A7Xk2npkSneDU0NAahKRES6gs+zlh544AH69u3Ld7/7XQzD4K9//Wub52mvJQk15wxIBmBnfhXF1Q2kxjpOPCk+CyITob4cinZB5hkBrVFERE5Np/ZaMgyDbdvaH0OgvZYkFKXEOhiaEcfO/CpW7ivhqjOyTjzJMDy9Mge+8KzwqyAjIhIWtNeS9AiTBqayM7+KL/e2E2TgeJAp2BrY4kRE5JRpRTvpEXzad+nrK/yKiEhY8CnIzJ49m5qaGp/fdM6cOZSWdjDVVSTAzh6QjMWA/cU1HC1vZ80j7bkkIhJ2fAoyf/zjH6mtrfX5TV944QXKy8tPtSaRLhcfaWdUn0Sgg16ZtGGex+p8qDnJjtkiIhISfBojY5omgwcP9nkwb2d6b0QCZdLAFDYdKufLvSVcP7bPiSc4YiGpP5R95VnhN+f8QJcoIiKd5FOQWbBgQaffOD09vdPXiPjTpIEpzF+6ly/2FGOaZtvBPH2kJ8gUKMiIiIQDn4LMjBkz/F2HiN+N75+Mw2Yhv7KevMJqBqfHnXhS+gjY+a7GyYiIhAnNWpIeI9JuZcIAz+ylz3YVtX2Sd+aSgoyISDhQkJEe5fxcz3YFy/LaCTIZozyPhdvBpY1QRURCnYKM9CgXDO4FwKr9pdQ1uk48IXmAZ6uCpnotjCciEgYUZKRHGZQWS2ZCJI1Nblbtb2OKtWFA77Gezw+vDWxxIiLSaZ0KMk6nE5vNxtat+p+qhCfDMDg/19Mrs2x3cdsn9RnneTyyLkBViYjIqepUkLHb7fTt2xeXq40ueZEwcX7z7aV2x8n0bg4y6pEREQl5nb619OCDD/LLX/5SWxBI2Dp3UCoWA/YUVre9XUHLraWSPKgrC2xxIiLSKT7vft3i+eefZ8+ePWRlZdGvXz9iYmJavb5+/fouK07EHxKi7YzOTmTDwXKW7S7ixrP7tj4hJgWScqBsPxxZD4MuCU6hIiJyUp0OMtdee60fyhAJrPNze3mCTF4bQQY842TK9nvGySjIiIiErE4HmUceecQfdYgE1PmDe/HHJXkszyumyeXGZv3GXdbe42DLmxonIyIS4jodZFqsW7eOHTt2ADBixAjGjBnTZUWJ+NvoPgnER9qorG9i0+EKxvZLan2Cd+bSWjBNz7RsEREJOZ0OMoWFhdx4440sXbqUxMREAMrLy7noootYuHAhvXr16uoaRbqczWrh3NxU3t+Sz7LdRScGmYxRYI2A2hLPJpLJOUGpU0REOtbpWUt33303VVVVbNu2jdLSUkpLS9m6dSuVlZXcc889/qhRxC+868m0NQ3b5ji+XYHWkxERCVmdDjIffvghL774IsOGDfMeGz58OC+88AIffPBBlxYn4k8t68lsOlRORW0b+yppPRkRkZDX6SDjdrux2+0nHLfb7bjd7i4pSiQQshKjyE2LxW3C8j1trPL79XEyIiISkjodZC6++GLuvfdejh496j125MgRfvrTn3LJJZqmKuGlpVfm87ZuL7UsjHdsMzQ1BrAqERHxVaeDzPPPP09lZSX9+/dn4MCBDBw4kJycHCorK3nuuef8UaOI35yXmwrAst1FmKbZ+sXkARCVDK4GKNgShOpERORkOj1rKTs7m/Xr1/Pxxx+zc+dOAIYNG8aUKVO6vDgRf5uQk0KEzcLRinr2FtUwKC32+IstO2HvWQyH1x3voRERkZDRqSDjdDqJiopi48aNXHrppVx66aX+qkskIKIirJzdP5nle4pZtruodZABzziZPYubx8ncHpQaRUSkfdr9Wnq88wd7bi+1PU5GM5dEREKZdr+WHu+85vVkVu4rpaHpGyG991mex9K9UKufeRGRUKPdr6XHG5oRR684B0VVDaz7qoxJg1KPvxidDMkDPUHmyHrI1VgwEZFQot2vpcczDIPzclN5a/0RPssrah1kwDNOpnSvZ5yMgoyISEjpVJBpamrCMAxuu+02+vTp46+aRALugsG9eGv9EZbtLmbO1G+8mHkmbH7Ds56MiIiElE6NkbHZbPz2t7+lqanJX/WIBMXk5l6YHccqKayqb/1i5mjPY76CjIhIqDmllX0/++wzf9QiEjSpsQ5G9o4HYHneN7YraNk8suIQ1JQEuDIREelIp8fITJ06lQceeIAtW7YwduzYEwb7XnPNNV1WnEggnZfbi61HKvk8r5hvn/W1W6eR8Z5Vfkv3Qf4mGHhx8IoUEZFWOh1k7rzzTgDmzZt3wmuGYWiNGQlb5+f2Yv7SvXyeV4TbbWKxGMdfzBztCTLHFGRERELJKe1+3d6HQoyEs7H9koiOsFJc3ciO/MrWL7aMkzm2KfCFiYhIuzodZILpqaeewjAMZs2aFexSpBuKsFmYOCAFgGW7vzFORkFGRCQk+RxkrrjiCioqKrzPn3rqKcrLy73PS0pKGD58eJcW93Vr1qzhP//zPznjjDP89jVEvr4bdisZzUGmdB/Uf6O3RkREgsbnIPPRRx/R0NDgff6b3/ym1TYFTU1N7Nq1q2ura1ZdXc306dP585//TFJSkl++hgjA+YM92xWsO1BGTcPXlhmISYGEbM/n+VuCUJmIiLTF58G+pml2+NyfZs6cyZVXXsmUKVP49a9/3eG5DQ0NrQJXZaXnf89OpxOn09llNbW8V1e+p7QvUO3dJyGCPklRHC6rY3leIRcP6eV9zZo+EkvFIVxH1uPufbZf6wgm/WwHjto6cNTWgdNVbe3r9Z2etRRoCxcuZP369axZs8an8+fOnctjjz12wvF///vfREdHd3V5LF68uMvfU9oXiPbuF2HhMBb+vngd9Xvd3uODK6MYBhxd+wHri/v6vY5g08924KitA0dtHTin29a1tbU+nedzkDEMA8MwTjjmT4cOHeLee+9l8eLFREZG+nTNnDlzmD17tvd5ZWUl2dnZfOtb3yI+Pr7LanM6nSxevJhLL70Uu93eZe8rbQtkezt2FPLFaxs54IzhiivO8x438qzwv2/Rx1ZCxhVX+LWGYNLPduCorQNHbR04XdXWLXdUTqZTt5ZuvfVWHA4HAPX19dxxxx3eBfG+fjunq6xbt47CwkLOOuss7zGXy8WyZct4/vnnaWhowGq1trrG4XB4a/w6u93ulx9ef72vtC0Q7X3ukHRsFoODpXUcqWikf2rzoo99xgJgFO/Gbjohout7+EKJfrYDR20dOGrrwDndtvb1Wp+DzIwZM1o9v+WWW0445/vf/76vb+eTSy65hC1bWg+s/MEPfsDQoUO5//77TwgxIl0h1mFjXP8kVu4r5bPdRceDTFwGxKRBTSEUbvfsii0iIkHlc5BZsGCBP+toU1xcHCNHjmx1LCYmhpSUlBOOi3SlCwaneYPMjEn9PQcNw7OezJ7FcGyjgoyISAgIqwXxRALlguZp2Cv2ltDQ9LUVqzOb1zHSwngiIiEh5GctfdPSpUuDXYL0AMMy4+gV56CoqoG1X5UxeZBnoTyt8CsiElrUIyPSBsMwvL0yn319ld+WIFOwHZoag1CZiIh8nYKMSDtaVvn9bNfXgkxiP4hMALcTinYGqTIREWmhICPSjvMGpWIYsKugimMVdZ6DLQN+QbeXRERCgIKMSDuSYiIY3ScR+MYmkgoyIiIhQ0FGpANtj5M50/OoICMiEnQKMiIduKB508jlecU0uZr3XcponoKdvwVc2oBORCSYFGREOjC6TyIJUXYq65vYdLjcczBlEEQlQ1MdHFkf1PpERHo6BRmRDlgtBufmetaQ+ff2As9BiwVyzvd8vm9pcAoTERFAQUbkpK4alQnAOxuP4nabnoMDLvQ8KsiIiASVgozISVw0NI24SBvHKupZtb/Uc7AlyBxeDQ3VQatNRKSnU5AROYlIu5Urm3tl3t5wxHMwOcezOJ67CQ58GcTqRER6NgUZER9MO7M3AO9vOUa9s3kTyYEXeR51e0lEJGgUZER8MCEnmcyESKoamvh0Z6HnoMbJiIgEnYKMiA8sFsPbK7Oo5fZS//MBAwq3QVVB8IoTEenBFGREfHTdGE+Q+XRXIeW1jRCTApnNi+Pt/yyIlYmI9FwKMiI+GpIRx7DMeJwuk/e2HPMc1O0lEZGgUpAR6YTrxmQB8K8NRz0Hvh5kTDMoNYmI9GQKMiKdcM3o3hgGrP6qlEOltdB3IlgdUHkESvYEuzwRkR5HQUakEzISIpk4IAWAdzYdBXsU9J3geVG3l0REAk5BRqSTrh1zfPaSaZowQOvJiIgEi4KMSCdNHZlBpN3CnsJq1h8sOz5OZv/n4GoKam0iIj2NgoxIJ8VF2rlmtGfQ76srD0LmaIhMhIYKOLYxqLWJiPQ0CjIip2D6hH4AvLvlGGV1Lsg53/PC7g+DWJWISM+jICNyCs7ok8DI3vE0Nrn5x7rDMHya54WVL2mVXxGRAFKQETkFhmF4e2VeW30Q9/DrIOssaKyCTx4PcnUiIj2HgozIKbpmdBZxDhv7i2v4cl8ZTH3a88KGV+HohuAWJyLSQyjIiJyiGIeN687yTMV+ddUByD4bRt0AmPDBA1rpV0QkABRkRE5Dy+2lf28voKCyHqY8BvZoOLQStr0V5OpERLo/BRmR0zAkI47x/ZNwuU3eWHMIEnrD5FmeF//9MDTWBrU+EZHuTkFG5DS19Mq8vvogTS43TLobErKh8jB8+VyQqxMR6d4UZERO0+UjM0iKtnOsop4lOwshIhoufczz4vJ5sH9ZcAsUEenGFGRETlOk3cp3x2cD8Ku3t3K0vA5GfBtyL4Omevj7d2Dn+0GuUkSke1KQEekCd100iCHpcRRVNfDD/15LdaMLvvs3GHoVuBrgjVtg0xvBLlNEpNtRkBHpAnGRdv5y6zhSYx3sOFbJPa9vwGV1wA3/DaNvAtMFi26HVS8Hu1QRkW5FQUaki/RJiua/ZozDYbPwyc5Cfv3edrDaYNqLMOEOz0kf3AfLfhfcQkVEuhEFGZEudGZ2IvO+eyYAC774ir+t+AosFrj8KbjgAc9JnzwBS58KWo0iIt2JgoxIF7vyjEzuu2wIAI+8s40/L9uHCXDRHJjyqOekpXPhkye1+q+IyGlSkBHxgzsvHMiMif0wTXjy/R3MeWsLjU1uOPen8K1fe05a9gwseVxhRkTkNCjIiPiBYRg8es0IHrpqOBYDFq45xIxXVlNe2+hZMO+yuZ4Tl8+DD+dAbWn7b2aa4HIGpnARkTCjICPiJ4Zh8MNzc/ivGeOIibCyYl8J1734JXuLqmHinTD1t54TV82H3w6Cv02DNX+Bqnwo3Alr/gv+cRv8fijMzYbt/wruNyQiEoJswS5ApLu7eGg6/7xzEj/861r2F9dw9XPLeWLaSK6fcDtEJcHyP0DhNti31PPx3uy23+jNH8D1/wUjvx3I8kVEQpp6ZEQCYGhGPG/PnMw5A5KpbXTxszc38dM3NlI95Dq480u4e71nIHDvsZ4LbFGQcz5c+EuY8S6MvtmzFs0/fwib3wzq9yIiEkrUIyMSIL3iHLz6H+fw4qd7eHZJHos2HGH9wTKeu2kMZ/QZ6BkIfO5Poa4M7DFgizh+cb9JnmncG/7uWVjPdMHoG4P3zYiIhAgFGZEAsloM7r4kl4kDU7h34UYOlNRyzfNfMLZfEledkcmVozJJi0868UKLFa5+DgwrrP9vWHQH7HofGmuhrtQzWNjtghHXehbfS+gd8O9NpDNM0ySvsJroCCtZCVFYLMaJJ7maYP9S6DMeIhMCXqOEBwUZkSAY1z+Z9+85jwff3sJ7W46x7kAZ6w6U8fi725mQk8yUYelMHJjCsIz44//AWyxw1bNgscHav7Q9+PfLP8HKFz2bVk66CzJHd6oul9tkw8FySuo9v2jaVLofVr8MmWfCqBs8dXWVY5tg00I482bIGNV17yshxelyc8/rG/hgaz4ADpuFfinR9E+J4eycZL43sR8Oi+Hpfdz6T0geCN9bBEn92n5DlxOs9gB+Bz1UU4NnQkLGKMg5L9jVeCnIiARJQrSd528+i19V1PP+lmO8u/ko6w+Ws3JfKSv3eaZjJ0bbOScnhTP7JuKwWTAAI+lu+o8eSh9LCX2ysnDEpUJUMtQWw8qX4MBy2PK/sOV/aYzJwh4RgWFYPL05VjvEZXp+IST2hcR+EJeB03CwOK+C/1lbQF6ZSTHx/OeeZYzrn8zYfkmcnZPMiCQXxue/h1X/Ce7m6eCr/xMufxqyx59eYzQ1wGdPw/JnPbfN1vwFrvgtnPV9MNr4n3ow1ZVB3scw6BKITu7a966vgJK9kDUm9L7vLvL1EGO1GFgMaGhys7ugmt0F1fx7ewGvrz7I/2S/S9b2f3ouKt0Lf/mWJ8ykD//am9V7ljBY/ixknw3Tnoek/sH4trq/ujJYeIvn3xcMmPo0TPhxsKsCFGREgi4jIZLbzs3htnNzOFxWy4db8/liTzGr95dSXuvkw235fLgt/xtX9QZ6Y7canJmdyMSBKQzN6M/WzN9TUr2Sc4veYKqxkoiao1DzjUsLt59Qgx24ovmDSCgzY9lUP5BNOwaybNtAjhjH6Gt/m3iqATCzz8Eo2ApH1sFfpsAZN8IlD0NMKribmj9cEBHbeqxPWw6vhbfvhOJdnufJA6B0H/zfPXBwJVz5e4iI7vg9TNPz0YneoSaXm33FNRwpr6OoqoHi6gaKqxqpqHNisxjYbQZ2q4UIq4UBvWK4fEQGCXv+BR/NgZoiz4yzKY/CmO93Ta/Uzvfg/2ZBTaGnJ+2C+2HIFacWaJoaoWSP55d6e21nmp7gFJkQsNDkdLm5d6EnxERYLfzn98dy3qBUjpbXs6+4mt0FVby8bD/nlf6TrKq/AVB74eNEb3sdinbAgsvh5jeh7wTYvwze/ann+wT46nOYP9mz4OTYWwMaBJ0uN3ZrN547U34QXr0BinZ6eoTdTfDBL6DsgKe9u7JX9hQYZrv9x91DZWUlCQkJVFRUEB8f32Xv63Q6ef/997niiiuw29Wl6W89sb2dLjdbjlSwYm8JuwuqcJvgNk0wPa9tO1rJkfK6dq8fElNLkrOABmcTFtxYMImxOsk0Ssk0C8g2isg2ikiiikijkWjDSazVic1Vj0Hb/yzsdGfzm6abyYubwEV9TKZX/ZURhf/X4fdhRqdgxGVCXAZEp4JhAdMNmNBY4xnrY7ohJg3zyt/TNPgKzOXPYv/sSQzTjTNlKAWTH6eqwU19dRkNNeWYteWkuQvo5TxKbO1hLOUHPD052RM8s71yLvD0alhtmPWVVBQdofjYAfJLK/iyLpvV+bDtaCV1Tpe3zkgaGG/ZxUDjKHvNLLa4cygnDoC+RgFP2hdwnmWz53uyRWI01XsuzDoLrvzd8RlnpomzppSqokNUlBZSVV5KdWUp9dVluCyRJI+4mBEjziDSbsXpdLL4nf/lcvMTLFv/cWLjpY+CC37h6W2oKYKaYs9HYzXE9z7es2aPau4pWgy73sfM+xijsQrTYsfIOhP6TvQMGLdHw5G1cGS9J0BW53t65UZcC8Ovbb8nyFkPh9fAV8vhwBeeWnIvhZHXe24xfv2apgbP+1cc8nzNhD6ewy439y7cyHtbjnlCzPfGctHQtBO+VM3Gt4h6+zYsmDztvJGFjusZlezmgbJHGO7aST0RrLKM4QL3KgCqI1LZPewuco68S1LxWgCq+lxAySW/IyGtPwlRdiwW4/i/IVOnYscJ9ZXgrAVbJKY9itJGK/k1JpW1dTRWldNUU4KrtgyzoZqoxF4k9OpDr/TepCXEUF7nZPX+UlbtK2HV/lJ25leRFG1nYK9YBvaKZVBaLBkJkVgMAwMTm7MKh7OcqPgUkpLTSI2LJDHajtHSbk2NULoXV9EumqpKaKqvwtVQg6u+BrOpHsMRjyU6EWt0IvboBCx2B7jdnp95043hasRSV4qlrsTz81Fb7OmBTcz2/PkmZHt+TlIGnfAfC9M0qahzkl9ZT35FPaU1jd6PstpGUqt28uPDDxDrLKEuMp3dU14hs2g5aas8i3q6h16N5fo/e34GW35cuujfa19/fyvInKKe+Is1mNTebTtYUsuXe4v5cm8Je4uqGZYZz9n9kxmfk0z/lGgaXW5W7ivl4+0FfLyjgGMV9a2uNwzISYnhP84bwPVje+OwWXHW1/Dlopc5t38U1vyNcHgtbtNke84M/qt6Eh/vLKG6ocn7HmcYe3nY/j+Ms+w+5e/jbfd5POb8HmVmrPfYOZbtPGd/jl5GxSm9Z70Rids0iObEsLfHncVa92C2WYYwKLqW8eZmBjdux2a2XkG5IiKTgqgc+lWsxUEjDaad55quZaHtGr4fsZTbnK8TSy1uDHZYhxLrrqCXu4Roo6HD2g6aaeyKGYul11DGHnyFRLMCFxb+x7iGl+qncKv9Y26xfERsG7W3pdKaTIyrHCvu49+/aSfS6NyK0EdIY4NlONEWN7FWJzFGI7HUktWwF7vZ2OY1hRF92JxwMbhd9K/ZTN/6nURw/OvutQ9mfcy5fMIElubb6W8tZu7FCZwZWwnVBZ4eo8gEiEyEpnp4dza4Gigcegu3HL2B3YWeLsVIGphvf5aLrJsAcJsGr7ou4ZmmG6kiGgM3t1k/5D7bG0QaTtymQQN2nNhoMmy4DDsRZgOx1LZqp69zmwYWo/1fiW7ToJQ4KswY6nBQRwR1poMGIrDgxk4TNlzYDBcx1JNsVJFCBRHG8cBcazo4aqZwjFRclgj6cZRsMx+b0XZNXanJsHPInsNeey47jYEcc8YQV3eY3u5j9DMKyDYKcWKjghgqzBgqieEyyxpijAZ2uLP5QeMvyCcFgGssX/Jb+0s4jCbWm7nsv+S/uP78MwEFmS6nINM9qL1Pn2maHC6rw2IxiLZbiYqwesbdfON/4Cdr63qni5X7PL1Eewqr2VtUw57Casy6ckwMLDYbNpsNm9WG0VBJoquUdKOMNKOMZKoAcGPg9oz4YYs7h9XmsDZr7kUZcyNe4UzLHmqNGOqtsTTaYnHaYjnkSmZrXQq7G1M4YKZjw8VEyzYmWbZzjmU7icbxe2rVZiQlRjJ2q0GW60j7jRTf2zOQsXi35/bW19T0Ppf/SbmHBTutFFQ2NNdXzgP217jeuvyEtyox46g24qi3xuCyx+KOiCfKWUq/uu3YcLU6d687k585f8JGc5D3WALV3Gb7gB9YPyKGOkqJo8RMoNSMow4HWUYJ2UYhscbxcLrL3YeP3Wex2DWOTeYA+hhFnG3sYrxlJ+Mtu4igiU3mADa6B7HRPZC9ZhbnWHZwpXUVF1s2dBjACsxEVrmHsdI9nHIzhiusq7nEsp4o48SAU2TGc9RMZZSxv8Ng0K6hV8F3/0aj2+CLPcU0utxE2a1E29z0X/sbHKU7WZ97D5sZzKGyWg6X1VHT0ERDk5u0xoPcV/dHRpkdB+sm00IdDhw0tgoaLWqNaGqtcTRaIolqqiTeXdFuAPJFAw4ctN++lWYUe83eFJkJ1OKg3ojCaYnEZdiJNGuJdtcQa9YQb9Riw4UbA1dzX2uTaaWcGErNeEqJp8SMw46L3kax96OvUUC84Vsw/qZ9ceOYn/4oB2tsFFU1UFnvpKq+idGubfw54vckGLXk9b+Z3FvnAwoyXU5BpntQewfOqbS1aZo0utzYLZZW02hN06SyvonCynoKqxoor3USFWEhym4jxmElOsKKw2bFajGwWYzmRwt2m0GE1YLVYpwQtL75dUtrGtlXXENhZQNu08RtmrhcTcRW7Sc+Jopemf3ond6LSLvVc1FNCRxe7Rl/c2QdRCV6bkUNuAhSBh6/TVJfAcc2Q/4Wzy2c5vEqLrfJ7oIq6pwunE1umtwmEUVbiKvaiz2pD5Ep2cSkZhMXG4e1jSnFZn0lBVs/pXzrv4ko2MhWdw754++nb0YqfZKi6RXnoLHJTb3TRZ3TRV19I3VOFzVOqGlsorahiTqn5xeqgUmUs5z4hqNYYlKJyxxIenwkGfGRJEZHUFTVwKGyWg6V1nKotI46p4tYh5UYh83zEWEjKsJCpM1KFPWkHFtGRMV+as0Iatx2qt12qlx2SmMGURPbH7vNgs1iwWKA021CQzVZBUvpW7QUlzWS0tSxVPYajzsphwibFaOmkJQjH5N+5GPSildiNZs8PS+JfT0fcZnQVOdp6/oKqCuHrDNh6jOtblV0mmlCdQHOxnoqq2uoqK6hrKKSdVt3M2j0OTRFxNNoicIwLKTFO8iIsZIW7cbhbgBrBDjiwfqNIaRuF+6aEsqKDuFwVhNrcXpuTTnrPN9Dy2B6i83zaI+G6BSI6eUZO2aP8pxbeZTG0oPUFH5FY3017uRBmKmDscVnYbdZibR7/oPR1nR0l9ukoclFk7v1r23TDS7TxOX2fDhdbuqcLkprGimvbaS0xklFbSMpzqNk1e0kvXonyRXbcLiqMZJycKTlYus10HMbynRDfbnnz6K+3PPnNfqmNse6OV1u6o5ux/bZUzRNe4H4OM8UeQWZLqYg0z2ovQNHbR04PaqtG2s8g0SDtB5Mj2rrIAt0kNGsJRER8b+ImGBXIN1UN54vJiIiIt2dgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiErZAOMnPnzmX8+PHExcWRlpbGtddey65du4JdloiIiISIkA4yn332GTNnzmTlypUsXrwYp9PJt771LWpqaoJdmoiIiISAkN408sMPP2z1/K9//StpaWmsW7eO888/P0hViYiISKgI6SDzTRUVFQAkJye3e05DQwMNDQ3e55WVlYBnW3Gn09lltbS8V1e+p7RP7R04auvAUVsHjto6cLqqrX293jBN0zytrxQgbreba665hvLycpYvX97ueY8++iiPPfbYCcdfe+01oqOj/VmiiIiIdJHa2lpuvvlmKioqiI+Pb/e8sAkyP/nJT/jggw9Yvnw5ffr0afe8tnpksrOzKS4u7rAhOsvpdLJ48WIuvfRS7HZ7l72vtE3tHThq68BRWweO2jpwuqqtKysrSU1NPWmQCYtbS3fddRfvvvsuy5Yt6zDEADgcDhwOxwnH7Xa7X354/fW+0ja1d+CorQNHbR04auvAOd229vXakA4ypmly9913s2jRIpYuXUpOTk6wSxIREZEQEtJBZubMmbz22mv861//Ii4ujvz8fAASEhKIiooKcnUiIiISbCG9jsz8+fOpqKjgwgsvJDMz0/vxxhtvBLs0ERERCQEh3SMTJuOQRUREJEhCukdGREREpCMKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMKWgoyIiIiELQUZERERCVsKMiIiIhK2FGREREQkbCnIiIiISNhSkBEREZGwpSAjIiIiYUtBRkRERMJWWASZF154gf79+xMZGcmECRNYvXp1sEsSERGREBDyQeaNN95g9uzZPPLII6xfv57Ro0dz2WWXUVhYGOzSREREJMhCPsjMmzePH/3oR/zgBz9g+PDhvPTSS0RHR/PKK68EuzQREREJMluwC+hIY2Mj69atY86cOd5jFouFKVOmsGLFijavaWhooKGhwfu8oqICgNLSUpxOZ5fV5nQ6qa2tpaSkBLvd3mXvK21TeweO2jpw1NaBo7YOnK5q66qqKgBM0+zwvJAOMsXFxbhcLtLT01sdT09PZ+fOnW1eM3fuXB577LETjufk5PilRhEREfGfqqoqEhIS2n09pIPMqZgzZw6zZ8/2Pne73ZSWlpKSkoJhGF32dSorK8nOzubQoUPEx8d32ftK29TegaO2Dhy1deCorQOnq9raNE2qqqrIysrq8LyQDjKpqalYrVYKCgpaHS8oKCAjI6PNaxwOBw6Ho9WxxMREf5VIfHy8/lIEkNo7cNTWgaO2Dhy1deB0RVt31BPTIqQH+0ZERDB27FiWLFniPeZ2u1myZAkTJ04MYmUiIiISCkK6RwZg9uzZzJgxg3HjxnH22Wfz7LPPUlNTww9+8INglyYiIiJBFvJB5v/9v/9HUVERDz/8MPn5+Zx55pl8+OGHJwwADjSHw8Ejjzxywm0s8Q+1d+CorQNHbR04auvACXRbG+bJ5jWJiIiIhKiQHiMjIiIi0hEFGREREQlbCjIiIiISthRkREREJGwpyJyiF154gf79+xMZGcmECRNYvXp1sEsKe3PnzmX8+PHExcWRlpbGtddey65du1qdU19fz8yZM0lJSSE2Npbrr7/+hAUTpfOeeuopDMNg1qxZ3mNq665z5MgRbrnlFlJSUoiKimLUqFGsXbvW+7ppmjz88MNkZmYSFRXFlClTyMvLC2LF4cnlcvHQQw+Rk5NDVFQUAwcO5Iknnmi1V4/a+tQsW7aMq6++mqysLAzD4O233271ui/tWlpayvTp04mPjycxMZEf/vCHVFdXn35xpnTawoULzYiICPOVV14xt23bZv7oRz8yExMTzYKCgmCXFtYuu+wyc8GCBebWrVvNjRs3mldccYXZt29fs7q62nvOHXfcYWZnZ5tLliwx165da55zzjnmpEmTglh1+Fu9erXZv39/84wzzjDvvfde73G1ddcoLS01+/XrZ956663mqlWrzH379pkfffSRuWfPHu85Tz31lJmQkGC+/fbb5qZNm8xrrrnGzMnJMevq6oJYefh58sknzZSUFPPdd9819+/fb7755ptmbGys+cc//tF7jtr61Lz//vvmgw8+aL711lsmYC5atKjV67606+WXX26OHj3aXLlypfn555+bgwYNMm+66abTrk1B5hScffbZ5syZM73PXS6XmZWVZc6dOzeIVXU/hYWFJmB+9tlnpmmaZnl5uWm3280333zTe86OHTtMwFyxYkWwygxrVVVVZm5urrl48WLzggsu8AYZtXXXuf/++81zzz233dfdbreZkZFh/va3v/UeKy8vNx0Oh/n6668HosRu48orrzRvu+22Vse+/e1vm9OnTzdNU23dVb4ZZHxp1+3bt5uAuWbNGu85H3zwgWkYhnnkyJHTqke3ljqpsbGRdevWMWXKFO8xi8XClClTWLFiRRAr634qKioASE5OBmDdunU4nc5WbT906FD69u2rtj9FM2fO5Morr2zVpqC27krvvPMO48aN44YbbiAtLY0xY8bw5z//2fv6/v37yc/Pb9XWCQkJTJgwQW3dSZMmTWLJkiXs3r0bgE2bNrF8+XKmTp0KqK39xZd2XbFiBYmJiYwbN857zpQpU7BYLKxateq0vn7Ir+wbaoqLi3G5XCesLJyens7OnTuDVFX343a7mTVrFpMnT2bkyJEA5OfnExERccImoOnp6eTn5wehyvC2cOFC1q9fz5o1a054TW3ddfbt28f8+fOZPXs2v/zlL1mzZg333HMPERERzJgxw9uebf2borbunAceeIDKykqGDh2K1WrF5XLx5JNPMn36dAC1tZ/40q75+fmkpaW1et1ms5GcnHzaba8gIyFp5syZbN26leXLlwe7lG7p0KFD3HvvvSxevJjIyMhgl9Otud1uxo0bx29+8xsAxowZw9atW3nppZeYMWNGkKvrXv73f/+XV199lddee40RI0awceNGZs2aRVZWltq6G9OtpU5KTU3FarWeMHujoKCAjIyMIFXVvdx11128++67fPrpp/Tp08d7PCMjg8bGRsrLy1udr7bvvHXr1lFYWMhZZ52FzWbDZrPx2Wef8ac//QmbzUZ6errauotkZmYyfPjwVseGDRvGwYMHAbztqX9TTt99993HAw88wI033sioUaP43ve+x09/+lPmzp0LqK39xZd2zcjIoLCwsNXrTU1NlJaWnnbbK8h0UkREBGPHjmXJkiXeY263myVLljBx4sQgVhb+TNPkrrvuYtGiRXzyySfk5OS0en3s2LHY7fZWbb9r1y4OHjyotu+kSy65hC1btrBx40bvx7hx45g+fbr3c7V115g8efIJywjs3r2bfv36AZCTk0NGRkartq6srGTVqlVq606qra3FYmn9a81qteJ2uwG1tb/40q4TJ06kvLycdevWec/55JNPcLvdTJgw4fQKOK2hwj3UwoULTYfDYf71r381t2/fbt5+++1mYmKimZ+fH+zSwtpPfvITMyEhwVy6dKl57Ngx70dtba33nDvuuMPs27ev+cknn5hr1641J06caE6cODGIVXcfX5+1ZJpq666yevVq02azmU8++aSZl5dnvvrqq2Z0dLT597//3XvOU089ZSYmJpr/+te/zM2bN5vTpk3TlOBTMGPGDLN3797e6ddvvfWWmZqaav7iF7/wnqO2PjVVVVXmhg0bzA0bNpiAOW/ePHPDhg3mgQMHTNP0rV0vv/xyc8yYMeaqVavM5cuXm7m5uZp+HUzPPfec2bdvXzMiIsI8++yzzZUrVwa7pLAHtPmxYMEC7zl1dXXmnXfeaSYlJZnR0dHmddddZx47dix4RXcj3wwyauuu83//93/myJEjTYfDYQ4dOtR8+eWXW73udrvNhx56yExPTzcdDod5ySWXmLt27QpSteGrsrLSvPfee82+ffuakZGR5oABA8wHH3zQbGho8J6jtj41n376aZv/Ps+YMcM0Td/ataSkxLzpppvM2NhYMz4+3vzBD35gVlVVnXZthml+bclDERERkTCiMTIiIiISthRkREREJGwpyIiIiEjYUpARERGRsKUgIyIiImFLQUZERETCloKMiIiIhC0FGRHpcQzD4O233w52GSLSBRRkRCSgbr31VgzDOOHj8ssvD3ZpIhKGbMEuQER6nssvv5wFCxa0OuZwOIJUjYiEM/XIiEjAORwOMjIyWn0kJSUBnts+8+fPZ+rUqURFRTFgwAD+8Y9/tLp+y5YtXHzxxURFRZGSksLtt99OdXV1q3NeeeUVRowYgcPhIDMzk7vuuqvV68XFxVx33XVER0eTm5vLO++8499vWkT8QkFGRELOQw89xPXXX8+mTZuYPn06N954Izt27ACgpqaGyy67jKSkJNasWcObb77Jxx9/3CqozJ8/n5kzZ3L77bezZcsW3nnnHQYNGtTqazz22GN897vfZfPmzVxxxRVMnz6d0tLSgH6fItIFTnvbSRGRTpgxY4ZptVrNmJiYVh9PPvmkaZqeXdDvuOOOVtdMmDDB/MlPfmKapmm+/PLLZlJSklldXe19/b333jMtFouZn59vmqZpZmVlmQ8++GC7NQDmr371K+/z6upqEzA/+OCDLvs+RSQwNEZGRALuoosuYv78+a2OJScnez+fOHFiq9cmTpzIxo0bAdixYwejR48mJibG+/rkyZNxu93s2rULwzA4evQol1xySYc1nHHGGd7PY2JiiI+Pp7Cw8FS/JREJEgUZEQm4mJiYE271dJWoqCifzrPb7a2eG4aB2+32R0ki4kcaIyMiIWflypUnPB82bBgAw4YNY9OmTdTU1Hhf/+KLL7BYLAwZMoS4uDj69+/PkiVLAlqziASHemREJOAaGhrIz89vdcxms5GamgrAm2++ybhx4zj33HN59dVXWb16NX/5y18AmD59Oo888ggzZszg0UcfpaioiLvvvpvvfe97pKenA/Doo49yxx13kJaWxtSpU6mqquKLL77g7rvvDuw3KiJ+pyAjIgH34YcfkpmZ2erYkCFD2LlzJ+CZUbRw4ULuvPNOMjMzef311xk+fDgA0dHRfPTRR9x7772MHz+e6Ohorr/+eubNm+d9rxkzZlBfX88f/vAHfv7zn5Oamsp3vvOdwH2DIhIwhmmaZrCLEBFpYRgGixYt4tprrw12KSISBjRGRkRERMKWgoyIiIiELY2REZGQorvdItIZ6pERERGRsKUgIyIiImFLQUZERETCloKMiIiIhC0FGREREQlbCjIiIiISthRkREREJGwpyIiIiEjYUpARERGRsPX/Ablo81Oq6yKAAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NyN49hIWe_NH"
   },
   "source": [
    "後で使用するために、テスト用セットの結果を収集します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:22.050741Z",
     "iopub.status.busy": "2022-08-09T01:44:22.050126Z",
     "iopub.status.idle": "2022-08-09T01:44:22.129500Z",
     "shell.execute_reply": "2022-08-09T01:44:22.128697Z"
    },
    "id": "jNC3D1DGsGgK"
   },
   "outputs": [],
   "source": [
    "test_results['linear_model'] = linear_model.evaluate(\n",
    "    test_features, test_labels, verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "SmjdzxKzEu1-"
   },
   "source": [
    "## DNN 回帰"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DT_aHPsrzO1t"
   },
   "source": [
    "前のセクションでは、単一および複数の入力の線形モデルを実装しました。\n",
    "\n",
    "このセクションでは、単一入力および複数入力の DNN モデルを実装します。コードは基本的に同じですが、モデルが拡張されていくつかの「非表示」の非線形レイヤーが含まれる点が異なります。「非表示」とは、入力または出力に直接接続されていないことを意味します。\n",
    "\n",
    "コードは基本的に同じですが、モデルが拡張されていくつかの「非表示」の非線形レイヤーが含まれる点が異なります。「非表示」とは、入力または出力に直接接続されていないことを意味します。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6SWtkIjhrZwa"
   },
   "source": [
    "これらのモデルには、線形モデルよりも多少多くのレイヤーが含まれます。\n",
    "\n",
    "- 前と同じく正規化レイヤー。（単一入力モデルの場合は `horsepower_normalizer`、複数入力モデルの場合は `normalizer` を使用）。\n",
    "- `relu` 非線形性を使用する 2 つの非表示の非線形`Dense` レイヤー。\n",
    "- 線形単一出力レイヤー\n",
    "\n",
    "どちらも同じトレーニング手順を使用するため、`compile` メソッドは以下の `build_and_compile_model` 関数に含まれています。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:22.133924Z",
     "iopub.status.busy": "2022-08-09T01:44:22.133266Z",
     "iopub.status.idle": "2022-08-09T01:44:22.137728Z",
     "shell.execute_reply": "2022-08-09T01:44:22.137068Z"
    },
    "id": "c26juK7ZG8j-"
   },
   "outputs": [],
   "source": [
    "def build_and_compile_model(norm):\n",
    "  model = keras.Sequential([\n",
    "      norm,\n",
    "      layers.Dense(64, activation='relu'),\n",
    "      layers.Dense(64, activation='relu'),\n",
    "      layers.Dense(1)\n",
    "  ])\n",
    "\n",
    "  model.compile(loss='mean_absolute_error',\n",
    "                optimizer=tf.keras.optimizers.Adam(0.001))\n",
    "  return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6c51caebbc0d"
   },
   "source": [
    "### DNN と単一入力を使用した回帰"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xvu9gtxTZR5V"
   },
   "source": [
    "入力 `'Horsepower'`、正規化レイヤー `horsepower_normalizer`（前に定義）のみを使用して DNN モデルを作成します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:22.140978Z",
     "iopub.status.busy": "2022-08-09T01:44:22.140519Z",
     "iopub.status.idle": "2022-08-09T01:44:22.186965Z",
     "shell.execute_reply": "2022-08-09T01:44:22.186305Z"
    },
    "id": "cGbPb-PHGbhs"
   },
   "outputs": [],
   "source": [
    "dnn_horsepower_model = build_and_compile_model(horsepower_normalizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Sj49Og4YGULr"
   },
   "source": [
    "このモデルには、線形モデルよりも多少多くのトレーニング可能なレイヤーが含まれます。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:22.190920Z",
     "iopub.status.busy": "2022-08-09T01:44:22.190322Z",
     "iopub.status.idle": "2022-08-09T01:44:22.202547Z",
     "shell.execute_reply": "2022-08-09T01:44:22.201940Z"
    },
    "id": "ReAD0n6MsFK-"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Layer (type)                Output Shape              Param #   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " normalization_1 (Normalizat  (None, 1)                3         \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " ion)                                                            \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense_2 (Dense)             (None, 64)                128       \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense_3 (Dense)             (None, 64)                4160      \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense_4 (Dense)             (None, 1)                 65        \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total params: 4,356\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable params: 4,353\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Non-trainable params: 3\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "dnn_horsepower_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "0-qWCsh6DlyH"
   },
   "source": [
    "Keras `Model.fit` を使用してモデルをトレーニングします。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:22.210239Z",
     "iopub.status.busy": "2022-08-09T01:44:22.209701Z",
     "iopub.status.idle": "2022-08-09T01:44:27.116191Z",
     "shell.execute_reply": "2022-08-09T01:44:27.115199Z"
    },
    "id": "sD7qHCmNIOY0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 6.03 s, sys: 1.52 s, total: 7.55 s\n",
      "Wall time: 4.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "history = dnn_horsepower_model.fit(\n",
    "    train_features['Horsepower'],\n",
    "    train_labels,\n",
    "    validation_split=0.2,\n",
    "    verbose=0, epochs=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "dArGGxHxcKjN"
   },
   "source": [
    "このモデルは、単一入力の線形 `horsepower_model` よりもわずかに優れています。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:27.120163Z",
     "iopub.status.busy": "2022-08-09T01:44:27.119598Z",
     "iopub.status.idle": "2022-08-09T01:44:27.239583Z",
     "shell.execute_reply": "2022-08-09T01:44:27.238947Z"
    },
    "id": "NcF6UWjdCU8T"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TG1snlpR2QCK"
   },
   "source": [
    "`Horsepower` の関数として予測をプロットすると、このモデルが非表示のレイヤーにより提供される非線形性をどのように利用するかがわかります。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:27.243394Z",
     "iopub.status.busy": "2022-08-09T01:44:27.242761Z",
     "iopub.status.idle": "2022-08-09T01:44:27.368185Z",
     "shell.execute_reply": "2022-08-09T01:44:27.367449Z"
    },
    "id": "hPF53Rem14NS"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "1/8 [==>...........................] - ETA: 0s"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "8/8 [==============================] - 0s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "x = tf.linspace(0.0, 250, 251)\n",
    "y = dnn_horsepower_model.predict(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:27.371663Z",
     "iopub.status.busy": "2022-08-09T01:44:27.371411Z",
     "iopub.status.idle": "2022-08-09T01:44:27.495759Z",
     "shell.execute_reply": "2022-08-09T01:44:27.495029Z"
    },
    "id": "rsf9rD8I17Wq"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_horsepower(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "WxCJKIUpe4io"
   },
   "source": [
    "後で使用するために、テスト用セットの結果を収集します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:27.499392Z",
     "iopub.status.busy": "2022-08-09T01:44:27.498906Z",
     "iopub.status.idle": "2022-08-09T01:44:27.565281Z",
     "shell.execute_reply": "2022-08-09T01:44:27.564533Z"
    },
    "id": "bJjM0dU52XtN"
   },
   "outputs": [],
   "source": [
    "test_results['dnn_horsepower_model'] = dnn_horsepower_model.evaluate(\n",
    "    test_features['Horsepower'], test_labels,\n",
    "    verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "S_2Btebp2e64"
   },
   "source": [
    "### 完全モデル"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aKFtezDldLSf"
   },
   "source": [
    "すべての入力を使用してこのプロセスを繰り返すと、検証データセットの性能がわずかに向上します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:27.569557Z",
     "iopub.status.busy": "2022-08-09T01:44:27.569040Z",
     "iopub.status.idle": "2022-08-09T01:44:27.617621Z",
     "shell.execute_reply": "2022-08-09T01:44:27.617035Z"
    },
    "id": "c0mhscXh2k36"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Layer (type)                Output Shape              Param #   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " normalization (Normalizatio  (None, 9)                19        \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " n)                                                              \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense_5 (Dense)             (None, 64)                640       \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense_6 (Dense)             (None, 64)                4160      \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dense_7 (Dense)             (None, 1)                 65        \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total params: 4,884\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable params: 4,865\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Non-trainable params: 19\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "dnn_model = build_and_compile_model(normalizer)\n",
    "dnn_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:27.624656Z",
     "iopub.status.busy": "2022-08-09T01:44:27.624105Z",
     "iopub.status.idle": "2022-08-09T01:44:32.491309Z",
     "shell.execute_reply": "2022-08-09T01:44:32.490583Z"
    },
    "id": "CXDENACl2tuW"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 6.09 s, sys: 1.49 s, total: 7.58 s\n",
      "Wall time: 4.86 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "history = dnn_model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    validation_split=0.2,\n",
    "    verbose=0, epochs=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:32.494959Z",
     "iopub.status.busy": "2022-08-09T01:44:32.494336Z",
     "iopub.status.idle": "2022-08-09T01:44:32.608038Z",
     "shell.execute_reply": "2022-08-09T01:44:32.607374Z"
    },
    "id": "-9Dbj0fX23RQ"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hWoVYS34fJPZ"
   },
   "source": [
    "後で使用するために、テスト用セットの結果を収集します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:32.611849Z",
     "iopub.status.busy": "2022-08-09T01:44:32.611354Z",
     "iopub.status.idle": "2022-08-09T01:44:32.681478Z",
     "shell.execute_reply": "2022-08-09T01:44:32.680673Z"
    },
    "id": "-bZIa96W3c7K"
   },
   "outputs": [],
   "source": [
    "test_results['dnn_model'] = dnn_model.evaluate(test_features, test_labels, verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uiCucdPLfMkZ"
   },
   "source": [
    "## 性能"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rDf1xebEfWBw"
   },
   "source": [
    "すべてのモデルがトレーニングされたので、テスト用セットの性能を確認します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:32.685782Z",
     "iopub.status.busy": "2022-08-09T01:44:32.685174Z",
     "iopub.status.idle": "2022-08-09T01:44:32.692666Z",
     "shell.execute_reply": "2022-08-09T01:44:32.692094Z"
    },
    "id": "e5_ooufM5iH2"
   },
   "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>Mean absolute error [MPG]</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>horsepower_model</th>\n",
       "      <td>3.636112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linear_model</th>\n",
       "      <td>2.459979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dnn_horsepower_model</th>\n",
       "      <td>2.934350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dnn_model</th>\n",
       "      <td>1.679483</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Mean absolute error [MPG]\n",
       "horsepower_model                       3.636112\n",
       "linear_model                           2.459979\n",
       "dnn_horsepower_model                   2.934350\n",
       "dnn_model                              1.679483"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(test_results, index=['Mean absolute error [MPG]']).T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DABIVzsCf-QI"
   },
   "source": [
    "これらの結果は、トレーニング中に見られる検証エラーと一致します。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ft603OzXuEZC"
   },
   "source": [
    "### モデルを使った予測\n",
    "\n",
    "Keras `Model.predict` を使用して、テストセットの `dnn_model` で予測を行い、損失を確認します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:32.696193Z",
     "iopub.status.busy": "2022-08-09T01:44:32.695568Z",
     "iopub.status.idle": "2022-08-09T01:44:32.901751Z",
     "shell.execute_reply": "2022-08-09T01:44:32.900980Z"
    },
    "id": "Xe7RXH3N3CWU"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r",
      "1/3 [=========>....................] - ETA: 0s"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "3/3 [==============================] - 0s 2ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_predictions = dnn_model.predict(test_features).flatten()\n",
    "\n",
    "a = plt.axes(aspect='equal')\n",
    "plt.scatter(test_labels, test_predictions)\n",
    "plt.xlabel('True Values [MPG]')\n",
    "plt.ylabel('Predictions [MPG]')\n",
    "lims = [0, 50]\n",
    "plt.xlim(lims)\n",
    "plt.ylim(lims)\n",
    "_ = plt.plot(lims, lims)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "19wyogbOSU5t"
   },
   "source": [
    "モデルの予測精度は妥当です。\n",
    "\n",
    "次に、エラー分布を見てみましょう。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:32.905773Z",
     "iopub.status.busy": "2022-08-09T01:44:32.905086Z",
     "iopub.status.idle": "2022-08-09T01:44:33.032420Z",
     "shell.execute_reply": "2022-08-09T01:44:33.031767Z"
    },
    "id": "f-OHX4DiXd8x"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "error = test_predictions - test_labels\n",
    "plt.hist(error, bins=25)\n",
    "plt.xlabel('Prediction Error [MPG]')\n",
    "_ = plt.ylabel('Count')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KSyaHUfDT-mZ"
   },
   "source": [
    "モデルに満足している場合は、後で使用できるように保存します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:33.036186Z",
     "iopub.status.busy": "2022-08-09T01:44:33.035517Z",
     "iopub.status.idle": "2022-08-09T01:44:33.866305Z",
     "shell.execute_reply": "2022-08-09T01:44:33.865394Z"
    },
    "id": "4-WwLlmfT-mb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: dnn_model/assets\n"
     ]
    }
   ],
   "source": [
    "dnn_model.save('dnn_model')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Benlnl8UT-me"
   },
   "source": [
    "モデルを再度読み込むと、同じ出力が得られます。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:33.870579Z",
     "iopub.status.busy": "2022-08-09T01:44:33.869931Z",
     "iopub.status.idle": "2022-08-09T01:44:34.245573Z",
     "shell.execute_reply": "2022-08-09T01:44:34.244746Z"
    },
    "id": "dyyyj2zVT-mf"
   },
   "outputs": [],
   "source": [
    "reloaded = tf.keras.models.load_model('dnn_model')\n",
    "\n",
    "test_results['reloaded'] = reloaded.evaluate(\n",
    "    test_features, test_labels, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-08-09T01:44:34.249874Z",
     "iopub.status.busy": "2022-08-09T01:44:34.249128Z",
     "iopub.status.idle": "2022-08-09T01:44:34.256783Z",
     "shell.execute_reply": "2022-08-09T01:44:34.256240Z"
    },
    "id": "f_GchJ2tg-2o"
   },
   "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>Mean absolute error [MPG]</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>horsepower_model</th>\n",
       "      <td>3.636112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linear_model</th>\n",
       "      <td>2.459979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dnn_horsepower_model</th>\n",
       "      <td>2.934350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dnn_model</th>\n",
       "      <td>1.679483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>reloaded</th>\n",
       "      <td>1.679483</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Mean absolute error [MPG]\n",
       "horsepower_model                       3.636112\n",
       "linear_model                           2.459979\n",
       "dnn_horsepower_model                   2.934350\n",
       "dnn_model                              1.679483\n",
       "reloaded                               1.679483"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(test_results, index=['Mean absolute error [MPG]']).T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vgGQuV-yqYZH"
   },
   "source": [
    "## まとめ\n",
    "\n",
    "このノートブックでは、回帰問題を扱うためのテクニックをいくつか紹介しました。\n",
    "\n",
    "- <a>平均二乗誤差 (MSE)</a> (`tf.keras.losses.MeanSquaredError`) および <a>平均絶対誤差 (MAE)</a> (`tf.keras.losses.MeanAbsoluteError`) は回帰問題に使われる一般的な損失関数です。MAE は外れ値の影響を受けにくくなっています。分類問題には異なる損失関数が使われます。\n",
    "- 同様に、回帰問題に使われる評価指標も分類問題とは異なります。\n",
    "- 入力数値特徴量の範囲が異なっている場合、特徴量ごとにおなじ範囲に正規化するべきです。\n",
    "- 過適合は DNN モデルの一般的な問題ですが、このチュートリアルでは問題ではありませんでした。これに関する詳細については、オーバーフィットとアンダーフィットのチュートリアルを参照してください。"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "regression.ipynb",
   "toc_visible": true
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}