03. Surprise(전체) + SVD.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 통계학세미나 추천 시스템 코드(surprise 패키지 이용)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "출처: https://datascienceschool.net/view-notebook/fcd3550f11ac4537acec8d18136f2066/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "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>beer</th>\n",
       "      <th>user</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>47658</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>66674</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>48824</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>3.71</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>55939</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>68916</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.34</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    beer          user  rating\n",
       "0  47658  bluejacket74    4.97\n",
       "1  66674  bluejacket74    4.04\n",
       "2  48824  bluejacket74    3.71\n",
       "3  55939  bluejacket74    4.10\n",
       "4  68916  bluejacket74    4.34"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#데이터 불러옴\n",
    "import pandas as pd\n",
    "df = pd.read_csv('my_df.csv')\n",
    "df.head() #상품아이디 / 사용자아이디 / 평점\n",
    "\n",
    "#추천 시스템은 사용자 아이디와 상품 아이디라는 2개의 카테고리 입력과, 평점 출력을 가지는 예측 시스템임."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(101889, 3)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "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>beer</th>\n",
       "      <th>user</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [beer, user, rating]\n",
       "Index: []"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.duplicated()==True] #중복 값 없음. 데이터엔 이상 x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>beer</th>\n",
       "      <th>user</th>\n",
       "      <th>rating</th>\n",
       "      <th>userid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>47658</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.97</td>\n",
       "      <td>190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>66674</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.04</td>\n",
       "      <td>190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>48824</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>3.71</td>\n",
       "      <td>190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>55939</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.10</td>\n",
       "      <td>190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>68916</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.34</td>\n",
       "      <td>190</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    beer          user  rating  userid\n",
       "0  47658  bluejacket74    4.97     190\n",
       "1  66674  bluejacket74    4.04     190\n",
       "2  48824  bluejacket74    3.71     190\n",
       "3  55939  bluejacket74    4.10     190\n",
       "4  68916  bluejacket74    4.34     190"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "user_enc = LabelEncoder()\n",
    "df['userid'] = user_enc.fit_transform(df['user'].values)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "333"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.userid.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([190, 101,  37, 301,  48, 119, 186, 323,  18, 171,  53, 296,  85,\n",
       "       293, 130, 262, 201,  98, 131, 118, 283, 327, 177,  97, 266, 318,\n",
       "       213,  50, 153,  38, 157,  33, 252, 164, 208,  75, 315, 281, 185,\n",
       "       121, 255, 326, 224, 238, 199, 267, 217, 256, 196,  79,  63, 173,\n",
       "        71, 114, 308, 161, 146,  56, 163,   3, 314, 269, 222, 100, 155,\n",
       "       182, 319, 166,  28,  47, 247,  67,  24,  62, 332, 138, 309, 126,\n",
       "       249, 188, 212, 102, 270,  54,  49, 104, 200,  57, 329, 144, 242,\n",
       "       274, 307,   4, 257, 250,  70, 209, 264,  35,  15, 127, 116, 234,\n",
       "       311,  76, 304,  73, 278, 261, 288, 174, 302, 122,   8,  81, 205,\n",
       "       179, 312, 152, 292, 136, 202, 260,  21,   9, 282,  41, 134, 135,\n",
       "       221, 271,  59,  14, 154, 223,  94, 253, 147,  80,  29, 306, 330,\n",
       "       277, 187, 273, 106,  13, 294, 139, 246, 218, 178, 245, 298, 317,\n",
       "       113, 142, 141, 272,   2, 310,   0,  61, 192, 280,  16,  23, 112,\n",
       "       244, 299,  64, 239, 220,  32, 284, 145,  22,  84, 148,  74, 225,\n",
       "       124,  12, 243,  58,  89, 259, 287,  60, 105,  99,  77, 233,   5,\n",
       "       276,  19,  69, 128, 248, 110, 129, 159, 211,  25, 297,  90, 151,\n",
       "       170,  82,  51,  10, 291, 169,  92, 176, 123,  43, 235, 321, 180,\n",
       "        78, 132,  11, 203, 300, 275,  42, 149, 263, 322, 168,  44,  31,\n",
       "       103,   1, 279, 210, 158,  86, 295, 206, 226, 289, 331,   7, 198,\n",
       "       165,   6, 140, 125, 197, 204, 137,  93, 268, 181, 240,  36,  40,\n",
       "       133, 324, 219, 290,  45, 254, 258, 228, 117,  95,  55, 285, 230,\n",
       "       241, 156, 108, 189,  91, 115,  65, 316, 214, 286,  17, 215, 229,\n",
       "       150, 236, 195, 237,  68,  87, 191, 328,  46,  34, 107, 227,  52,\n",
       "       167, 231, 305, 184, 175, 325, 111, 160, 162, 193, 120, 320, 194,\n",
       "       251, 303, 265,  30,  26, 216,  66, 109,  39,  96, 172,  88, 207,\n",
       "       313,  20, 143, 183,  27, 232,  83,  72], dtype=int64)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.userid.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "333"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#유저를 이름 말고 숫자로 바꿔보는거 이런 방법도 있음\n",
    "'''''user_list = df.user.unique().tolist()\n",
    "    df['int_user']=df['user']\n",
    "    for n in range(0, 333):\n",
    "        ind = df[df.user==user_list[n]].index\n",
    "        df.loc[ind, 'int_user']=int(n)\n",
    "\n",
    "    df.head() \n",
    "''''''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(333, 962)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#피봇 테이블(x축: 상품, y축: 사용자) : 평점행렬 R\n",
    "df_table = df.set_index(['userid','beer']).unstack()\n",
    "df_table.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"21\" halign=\"left\">rating</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beer</th>\n",
       "      <th>6</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>33</th>\n",
       "      <th>34</th>\n",
       "      <th>39</th>\n",
       "      <th>59</th>\n",
       "      <th>61</th>\n",
       "      <th>63</th>\n",
       "      <th>65</th>\n",
       "      <th>...</th>\n",
       "      <th>226</th>\n",
       "      <th>228</th>\n",
       "      <th>229</th>\n",
       "      <th>232</th>\n",
       "      <th>236</th>\n",
       "      <th>246</th>\n",
       "      <th>248</th>\n",
       "      <th>276</th>\n",
       "      <th>279</th>\n",
       "      <th>283</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>57md</th>\n",
       "      <td>3.69</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.63</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>2.18</td>\n",
       "      <td>...</td>\n",
       "      <td>3.87</td>\n",
       "      <td>3.79</td>\n",
       "      <td>4.29</td>\n",
       "      <td>1.69</td>\n",
       "      <td>3.61</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.97</td>\n",
       "      <td>4.06</td>\n",
       "      <td>4.13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ALC82</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.73</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.9</td>\n",
       "      <td>2.25</td>\n",
       "      <td>...</td>\n",
       "      <td>4.28</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3</td>\n",
       "      <td></td>\n",
       "      <td>3.75</td>\n",
       "      <td></td>\n",
       "      <td>4.06</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AgentMunky</th>\n",
       "      <td>4.07</td>\n",
       "      <td>3.95</td>\n",
       "      <td>3.16</td>\n",
       "      <td>3.88</td>\n",
       "      <td>4.3</td>\n",
       "      <td></td>\n",
       "      <td>4.15</td>\n",
       "      <td></td>\n",
       "      <td>3.46</td>\n",
       "      <td>1.72</td>\n",
       "      <td>...</td>\n",
       "      <td>3.85</td>\n",
       "      <td>4.43</td>\n",
       "      <td>3.68</td>\n",
       "      <td>1.96</td>\n",
       "      <td>3.33</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.4</td>\n",
       "      <td></td>\n",
       "      <td>3.26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Alieniloquium</th>\n",
       "      <td></td>\n",
       "      <td>4.07</td>\n",
       "      <td></td>\n",
       "      <td>3.88</td>\n",
       "      <td>4.12</td>\n",
       "      <td>3.95</td>\n",
       "      <td>3.41</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.2</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Arbitrator</th>\n",
       "      <td></td>\n",
       "      <td>2.51</td>\n",
       "      <td>3.27</td>\n",
       "      <td>3.56</td>\n",
       "      <td>4.28</td>\n",
       "      <td>4.32</td>\n",
       "      <td>4.01</td>\n",
       "      <td>4</td>\n",
       "      <td>4.17</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.23</td>\n",
       "      <td>4.48</td>\n",
       "      <td>4.5</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.7</td>\n",
       "      <td></td>\n",
       "      <td>3.41</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ArrogantB</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.57</td>\n",
       "      <td>4.1</td>\n",
       "      <td>3.36</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>3.36</td>\n",
       "      <td>3.88</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.43</td>\n",
       "      <td>4.3</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BARFLYB</th>\n",
       "      <td></td>\n",
       "      <td>3.77</td>\n",
       "      <td></td>\n",
       "      <td>3.72</td>\n",
       "      <td>4.35</td>\n",
       "      <td></td>\n",
       "      <td>4.6</td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td>2.76</td>\n",
       "      <td>...</td>\n",
       "      <td>4.08</td>\n",
       "      <td>3.85</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.43</td>\n",
       "      <td></td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.92</td>\n",
       "      <td>2.93</td>\n",
       "      <td>3.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BB1313</th>\n",
       "      <td>3.6</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.05</td>\n",
       "      <td></td>\n",
       "      <td>3.64</td>\n",
       "      <td>3.58</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.42</td>\n",
       "      <td>4.26</td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.81</td>\n",
       "      <td></td>\n",
       "      <td>3.61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BEER88</th>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.13</td>\n",
       "      <td>3.5</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>4.25</td>\n",
       "      <td>3.96</td>\n",
       "      <td></td>\n",
       "      <td>2.5</td>\n",
       "      <td></td>\n",
       "      <td>2.46</td>\n",
       "      <td>3.21</td>\n",
       "      <td>3.75</td>\n",
       "      <td>3.53</td>\n",
       "      <td>4.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BEERMILER12</th>\n",
       "      <td></td>\n",
       "      <td>4.42</td>\n",
       "      <td>3.9</td>\n",
       "      <td>4.19</td>\n",
       "      <td>4.6</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.35</td>\n",
       "      <td>4.13</td>\n",
       "      <td>2.95</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>2.38</td>\n",
       "      <td></td>\n",
       "      <td>2.91</td>\n",
       "      <td></td>\n",
       "      <td>4.13</td>\n",
       "      <td>4</td>\n",
       "      <td>4.22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BEERchitect</th>\n",
       "      <td>3</td>\n",
       "      <td>4.08</td>\n",
       "      <td>3.93</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.95</td>\n",
       "      <td>4.21</td>\n",
       "      <td>3.83</td>\n",
       "      <td>4.15</td>\n",
       "      <td>3.53</td>\n",
       "      <td>2.18</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>4.35</td>\n",
       "      <td>4.28</td>\n",
       "      <td>1.76</td>\n",
       "      <td>4</td>\n",
       "      <td>2.98</td>\n",
       "      <td>4.45</td>\n",
       "      <td>4.42</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BMMillsy</th>\n",
       "      <td></td>\n",
       "      <td>4.31</td>\n",
       "      <td>3.73</td>\n",
       "      <td></td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.27</td>\n",
       "      <td>3.86</td>\n",
       "      <td></td>\n",
       "      <td>3.1</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.36</td>\n",
       "      <td>4.27</td>\n",
       "      <td>4.26</td>\n",
       "      <td></td>\n",
       "      <td>3.82</td>\n",
       "      <td></td>\n",
       "      <td>3.59</td>\n",
       "      <td>3.71</td>\n",
       "      <td>3.51</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BarryMFBurton</th>\n",
       "      <td>3.95</td>\n",
       "      <td>4.02</td>\n",
       "      <td></td>\n",
       "      <td>4.47</td>\n",
       "      <td>4.15</td>\n",
       "      <td>4.09</td>\n",
       "      <td></td>\n",
       "      <td>4.24</td>\n",
       "      <td>3.61</td>\n",
       "      <td>1.92</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td>3.85</td>\n",
       "      <td>3.72</td>\n",
       "      <td></td>\n",
       "      <td>3.77</td>\n",
       "      <td>2.56</td>\n",
       "      <td></td>\n",
       "      <td>3.95</td>\n",
       "      <td>3.77</td>\n",
       "      <td>4.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BeerFMAndy</th>\n",
       "      <td>3.86</td>\n",
       "      <td>4.74</td>\n",
       "      <td>3.97</td>\n",
       "      <td>4.34</td>\n",
       "      <td>4.53</td>\n",
       "      <td>4.15</td>\n",
       "      <td>4.09</td>\n",
       "      <td>4.26</td>\n",
       "      <td>4.21</td>\n",
       "      <td>1.78</td>\n",
       "      <td>...</td>\n",
       "      <td>4.78</td>\n",
       "      <td>4.89</td>\n",
       "      <td>4.36</td>\n",
       "      <td>2.36</td>\n",
       "      <td></td>\n",
       "      <td>3.26</td>\n",
       "      <td>3.8</td>\n",
       "      <td>4.63</td>\n",
       "      <td>4.35</td>\n",
       "      <td>4.84</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BeerLover99</th>\n",
       "      <td></td>\n",
       "      <td>4.32</td>\n",
       "      <td>3.83</td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.95</td>\n",
       "      <td>4.42</td>\n",
       "      <td>4.2</td>\n",
       "      <td></td>\n",
       "      <td>3.98</td>\n",
       "      <td>2.35</td>\n",
       "      <td>...</td>\n",
       "      <td>4.23</td>\n",
       "      <td>4.43</td>\n",
       "      <td>3.91</td>\n",
       "      <td>2</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.92</td>\n",
       "      <td>4.42</td>\n",
       "      <td>3.63</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Beerandraiderfan</th>\n",
       "      <td>2.63</td>\n",
       "      <td>3.44</td>\n",
       "      <td>2.5</td>\n",
       "      <td>3.23</td>\n",
       "      <td>2.38</td>\n",
       "      <td>4.27</td>\n",
       "      <td>3.06</td>\n",
       "      <td>4</td>\n",
       "      <td>3.73</td>\n",
       "      <td>2.4</td>\n",
       "      <td>...</td>\n",
       "      <td>4.22</td>\n",
       "      <td>3.43</td>\n",
       "      <td>3.58</td>\n",
       "      <td>1.57</td>\n",
       "      <td>3.28</td>\n",
       "      <td>2.97</td>\n",
       "      <td>3.82</td>\n",
       "      <td>4.83</td>\n",
       "      <td>4.22</td>\n",
       "      <td>4.41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Beginner2</th>\n",
       "      <td>3.86</td>\n",
       "      <td>4.1</td>\n",
       "      <td>3.76</td>\n",
       "      <td>4.13</td>\n",
       "      <td>4.38</td>\n",
       "      <td>4.29</td>\n",
       "      <td>4.31</td>\n",
       "      <td>4.18</td>\n",
       "      <td>3.95</td>\n",
       "      <td>2.72</td>\n",
       "      <td>...</td>\n",
       "      <td>4.12</td>\n",
       "      <td>4</td>\n",
       "      <td>4.16</td>\n",
       "      <td>2.25</td>\n",
       "      <td>3.99</td>\n",
       "      <td>3.13</td>\n",
       "      <td>4</td>\n",
       "      <td>3.88</td>\n",
       "      <td>3.75</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bierguy5</th>\n",
       "      <td></td>\n",
       "      <td>3.9</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.25</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.78</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.52</td>\n",
       "      <td>4.13</td>\n",
       "      <td>4.38</td>\n",
       "      <td>2.4</td>\n",
       "      <td></td>\n",
       "      <td>2.35</td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BigMike</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.64</td>\n",
       "      <td>1.4</td>\n",
       "      <td>...</td>\n",
       "      <td>3.6</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>1.53</td>\n",
       "      <td></td>\n",
       "      <td>2.25</td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>3.59</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Biggiekrock627</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>4.49</td>\n",
       "      <td>3.97</td>\n",
       "      <td>3.68</td>\n",
       "      <td></td>\n",
       "      <td>4.16</td>\n",
       "      <td>1.7</td>\n",
       "      <td>...</td>\n",
       "      <td>3.98</td>\n",
       "      <td>4.13</td>\n",
       "      <td>3.69</td>\n",
       "      <td>1.08</td>\n",
       "      <td></td>\n",
       "      <td>2.38</td>\n",
       "      <td>3.65</td>\n",
       "      <td>3.77</td>\n",
       "      <td></td>\n",
       "      <td>3.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bighuge</th>\n",
       "      <td>3</td>\n",
       "      <td>4.59</td>\n",
       "      <td>3.52</td>\n",
       "      <td>4.22</td>\n",
       "      <td>4.33</td>\n",
       "      <td>3.15</td>\n",
       "      <td>4.52</td>\n",
       "      <td>3.11</td>\n",
       "      <td>4.38</td>\n",
       "      <td>1.5</td>\n",
       "      <td>...</td>\n",
       "      <td>4.5</td>\n",
       "      <td>5</td>\n",
       "      <td>4.53</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.28</td>\n",
       "      <td>4.47</td>\n",
       "      <td>4.33</td>\n",
       "      <td>4.15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Billolick</th>\n",
       "      <td>3.41</td>\n",
       "      <td>4.02</td>\n",
       "      <td>4.3</td>\n",
       "      <td>4.19</td>\n",
       "      <td>4.08</td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.15</td>\n",
       "      <td>3.98</td>\n",
       "      <td>2.28</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>2.68</td>\n",
       "      <td>3.87</td>\n",
       "      <td>2.9</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.27</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bitterbill</th>\n",
       "      <td></td>\n",
       "      <td>3.9</td>\n",
       "      <td></td>\n",
       "      <td>4.23</td>\n",
       "      <td>4.22</td>\n",
       "      <td>4.65</td>\n",
       "      <td>4.45</td>\n",
       "      <td>4.47</td>\n",
       "      <td>4.13</td>\n",
       "      <td>3.38</td>\n",
       "      <td>...</td>\n",
       "      <td>4.68</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>1.31</td>\n",
       "      <td>4.32</td>\n",
       "      <td>3.45</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.18</td>\n",
       "      <td>4.13</td>\n",
       "      <td>4.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bouleboubier</th>\n",
       "      <td>4.12</td>\n",
       "      <td>4.17</td>\n",
       "      <td>3.95</td>\n",
       "      <td>4.54</td>\n",
       "      <td>4.92</td>\n",
       "      <td></td>\n",
       "      <td>3.88</td>\n",
       "      <td>4.85</td>\n",
       "      <td>3.93</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.15</td>\n",
       "      <td>4.32</td>\n",
       "      <td>2.87</td>\n",
       "      <td>3.89</td>\n",
       "      <td>3.92</td>\n",
       "      <td>4.3</td>\n",
       "      <td>4.15</td>\n",
       "      <td>3.9</td>\n",
       "      <td>3.96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Brad007</th>\n",
       "      <td>3.41</td>\n",
       "      <td>4.13</td>\n",
       "      <td>4.85</td>\n",
       "      <td>3.93</td>\n",
       "      <td>4.82</td>\n",
       "      <td></td>\n",
       "      <td>3.88</td>\n",
       "      <td>2.81</td>\n",
       "      <td>3.12</td>\n",
       "      <td>2.1</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td>3.88</td>\n",
       "      <td>4.38</td>\n",
       "      <td>1</td>\n",
       "      <td>3.55</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.85</td>\n",
       "      <td>3.2</td>\n",
       "      <td>3.73</td>\n",
       "      <td>3.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Brenden</th>\n",
       "      <td>3.6</td>\n",
       "      <td>4.4</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.37</td>\n",
       "      <td>4.73</td>\n",
       "      <td>4.47</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.03</td>\n",
       "      <td>2.73</td>\n",
       "      <td>...</td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.3</td>\n",
       "      <td>4.2</td>\n",
       "      <td>1.52</td>\n",
       "      <td>3.68</td>\n",
       "      <td>2.85</td>\n",
       "      <td>3.63</td>\n",
       "      <td>4.12</td>\n",
       "      <td>3.98</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Brent</th>\n",
       "      <td>3.06</td>\n",
       "      <td>4.06</td>\n",
       "      <td>4</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4</td>\n",
       "      <td>4.24</td>\n",
       "      <td></td>\n",
       "      <td>4.8</td>\n",
       "      <td>4</td>\n",
       "      <td>1.1</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>1.84</td>\n",
       "      <td></td>\n",
       "      <td>1.82</td>\n",
       "      <td>4.4</td>\n",
       "      <td>4.4</td>\n",
       "      <td></td>\n",
       "      <td>4.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BretSikkink</th>\n",
       "      <td></td>\n",
       "      <td>3.88</td>\n",
       "      <td>4.05</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.85</td>\n",
       "      <td>4.38</td>\n",
       "      <td>4.18</td>\n",
       "      <td>4.23</td>\n",
       "      <td>3.63</td>\n",
       "      <td>2.97</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>2.35</td>\n",
       "      <td>3.43</td>\n",
       "      <td>2.72</td>\n",
       "      <td>3.85</td>\n",
       "      <td>4.1</td>\n",
       "      <td>3.9</td>\n",
       "      <td>4.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BucannonXC5</th>\n",
       "      <td>3.76</td>\n",
       "      <td>4.16</td>\n",
       "      <td></td>\n",
       "      <td>4.01</td>\n",
       "      <td>4.27</td>\n",
       "      <td></td>\n",
       "      <td>4.05</td>\n",
       "      <td>3.76</td>\n",
       "      <td>3.61</td>\n",
       "      <td>2.54</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>2.56</td>\n",
       "      <td></td>\n",
       "      <td>1.68</td>\n",
       "      <td>3.9</td>\n",
       "      <td>4.18</td>\n",
       "      <td>3.8</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Buck89</th>\n",
       "      <td></td>\n",
       "      <td>4.34</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.35</td>\n",
       "      <td>4.24</td>\n",
       "      <td>4.28</td>\n",
       "      <td>3.94</td>\n",
       "      <td>3.88</td>\n",
       "      <td>2.29</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>1.85</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.16</td>\n",
       "      <td></td>\n",
       "      <td>4.34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BuckeyeNation</th>\n",
       "      <td>2.93</td>\n",
       "      <td>3.92</td>\n",
       "      <td>3.42</td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.05</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.57</td>\n",
       "      <td>4.35</td>\n",
       "      <td>3.65</td>\n",
       "      <td>1.62</td>\n",
       "      <td>...</td>\n",
       "      <td>4.83</td>\n",
       "      <td>4.57</td>\n",
       "      <td>4.65</td>\n",
       "      <td>1.88</td>\n",
       "      <td>3.33</td>\n",
       "      <td>3</td>\n",
       "      <td>3.97</td>\n",
       "      <td>4</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CBrodbeck</th>\n",
       "      <td></td>\n",
       "      <td>4.13</td>\n",
       "      <td></td>\n",
       "      <td>3.94</td>\n",
       "      <td>4.18</td>\n",
       "      <td>4.13</td>\n",
       "      <td></td>\n",
       "      <td>4.04</td>\n",
       "      <td>3.77</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td>4.1</td>\n",
       "      <td></td>\n",
       "      <td>2.31</td>\n",
       "      <td></td>\n",
       "      <td>2.47</td>\n",
       "      <td>3.71</td>\n",
       "      <td>4.1</td>\n",
       "      <td></td>\n",
       "      <td>4.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CHickman</th>\n",
       "      <td>3.67</td>\n",
       "      <td>4.2</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.35</td>\n",
       "      <td>4.45</td>\n",
       "      <td></td>\n",
       "      <td>4.48</td>\n",
       "      <td>3.81</td>\n",
       "      <td>2.74</td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>1.81</td>\n",
       "      <td></td>\n",
       "      <td>3.17</td>\n",
       "      <td>3.98</td>\n",
       "      <td>4.67</td>\n",
       "      <td>4.39</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ChainGangGuy</th>\n",
       "      <td>3.23</td>\n",
       "      <td>4.29</td>\n",
       "      <td>3.53</td>\n",
       "      <td>3.78</td>\n",
       "      <td>4.02</td>\n",
       "      <td></td>\n",
       "      <td>3.25</td>\n",
       "      <td>3.68</td>\n",
       "      <td>3.53</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.15</td>\n",
       "      <td>3.93</td>\n",
       "      <td>4.08</td>\n",
       "      <td>1.96</td>\n",
       "      <td>3.18</td>\n",
       "      <td>2.68</td>\n",
       "      <td>3.3</td>\n",
       "      <td>3.52</td>\n",
       "      <td>3.03</td>\n",
       "      <td>3.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chaney</th>\n",
       "      <td></td>\n",
       "      <td>4.32</td>\n",
       "      <td>4</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.32</td>\n",
       "      <td>4.85</td>\n",
       "      <td>4.38</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.48</td>\n",
       "      <td>4.33</td>\n",
       "      <td>4.2</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CharlesK</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td>3.94</td>\n",
       "      <td>4.14</td>\n",
       "      <td>4.11</td>\n",
       "      <td>4</td>\n",
       "      <td>3.84</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>3.11</td>\n",
       "      <td>2.58</td>\n",
       "      <td>3.5</td>\n",
       "      <td>2.47</td>\n",
       "      <td></td>\n",
       "      <td>1.64</td>\n",
       "      <td>2.84</td>\n",
       "      <td>3.44</td>\n",
       "      <td>3.29</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chico1985</th>\n",
       "      <td></td>\n",
       "      <td>3.98</td>\n",
       "      <td></td>\n",
       "      <td>4.33</td>\n",
       "      <td>4.33</td>\n",
       "      <td>4.03</td>\n",
       "      <td>4.45</td>\n",
       "      <td></td>\n",
       "      <td>3.75</td>\n",
       "      <td>1.63</td>\n",
       "      <td>...</td>\n",
       "      <td>4.35</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>2.74</td>\n",
       "      <td></td>\n",
       "      <td>2.2</td>\n",
       "      <td>4.03</td>\n",
       "      <td>4.13</td>\n",
       "      <td></td>\n",
       "      <td>3.88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ColForbinBC</th>\n",
       "      <td></td>\n",
       "      <td>4.32</td>\n",
       "      <td>3.92</td>\n",
       "      <td>4.18</td>\n",
       "      <td>4.18</td>\n",
       "      <td>4.4</td>\n",
       "      <td>4.2</td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>1.88</td>\n",
       "      <td>...</td>\n",
       "      <td>4.32</td>\n",
       "      <td>4.1</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.98</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cpmitchno1</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.42</td>\n",
       "      <td>4</td>\n",
       "      <td>4.17</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.12</td>\n",
       "      <td></td>\n",
       "      <td>4.02</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.96</td>\n",
       "      <td></td>\n",
       "      <td>4.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CrellMoset</th>\n",
       "      <td>3.5</td>\n",
       "      <td>4.25</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4.4</td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.35</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.42</td>\n",
       "      <td>4.08</td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.01</td>\n",
       "      <td></td>\n",
       "      <td>3.92</td>\n",
       "      <td>3.83</td>\n",
       "      <td>4.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CuriousMonk</th>\n",
       "      <td>3</td>\n",
       "      <td>4.05</td>\n",
       "      <td></td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.47</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>3.8</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>3.82</td>\n",
       "      <td>3.65</td>\n",
       "      <td>4</td>\n",
       "      <td>2.93</td>\n",
       "      <td></td>\n",
       "      <td>2.6</td>\n",
       "      <td></td>\n",
       "      <td>3.75</td>\n",
       "      <td>3.75</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cyberkedi</th>\n",
       "      <td>4</td>\n",
       "      <td>4.41</td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.02</td>\n",
       "      <td></td>\n",
       "      <td>4.08</td>\n",
       "      <td>4.41</td>\n",
       "      <td>3.54</td>\n",
       "      <td>2.17</td>\n",
       "      <td>...</td>\n",
       "      <td>4.33</td>\n",
       "      <td>3.72</td>\n",
       "      <td></td>\n",
       "      <td>2.67</td>\n",
       "      <td>3.91</td>\n",
       "      <td>3.05</td>\n",
       "      <td></td>\n",
       "      <td>4.06</td>\n",
       "      <td>3.83</td>\n",
       "      <td>4.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DIM</th>\n",
       "      <td></td>\n",
       "      <td>4.37</td>\n",
       "      <td>4</td>\n",
       "      <td>3.8</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.85</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.47</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>3.76</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Damian</th>\n",
       "      <td></td>\n",
       "      <td>4.3</td>\n",
       "      <td></td>\n",
       "      <td>4.24</td>\n",
       "      <td>4.69</td>\n",
       "      <td>4.65</td>\n",
       "      <td>3.66</td>\n",
       "      <td>4.17</td>\n",
       "      <td>3.58</td>\n",
       "      <td>2.81</td>\n",
       "      <td>...</td>\n",
       "      <td>3.6</td>\n",
       "      <td>3.56</td>\n",
       "      <td></td>\n",
       "      <td>2.89</td>\n",
       "      <td></td>\n",
       "      <td>1.94</td>\n",
       "      <td>4.08</td>\n",
       "      <td>2.96</td>\n",
       "      <td></td>\n",
       "      <td>2.52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Daniellobo</th>\n",
       "      <td>3.38</td>\n",
       "      <td>4.07</td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>3.28</td>\n",
       "      <td></td>\n",
       "      <td>3.98</td>\n",
       "      <td>4.1</td>\n",
       "      <td>3.62</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td></td>\n",
       "      <td>4.21</td>\n",
       "      <td>4.13</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.98</td>\n",
       "      <td>3.83</td>\n",
       "      <td>3.88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Darkmagus82</th>\n",
       "      <td>3.85</td>\n",
       "      <td>4.14</td>\n",
       "      <td>3.75</td>\n",
       "      <td>4.09</td>\n",
       "      <td>4.37</td>\n",
       "      <td>4.03</td>\n",
       "      <td>3.89</td>\n",
       "      <td>3.9</td>\n",
       "      <td>3.87</td>\n",
       "      <td>2.85</td>\n",
       "      <td>...</td>\n",
       "      <td>4.34</td>\n",
       "      <td>3.83</td>\n",
       "      <td>4.11</td>\n",
       "      <td></td>\n",
       "      <td>3.93</td>\n",
       "      <td>2.21</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.96</td>\n",
       "      <td>3.77</td>\n",
       "      <td>4.12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DavoleBomb</th>\n",
       "      <td></td>\n",
       "      <td>3.7</td>\n",
       "      <td></td>\n",
       "      <td>4.33</td>\n",
       "      <td>4.52</td>\n",
       "      <td>4.2</td>\n",
       "      <td>4.39</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>1.89</td>\n",
       "      <td>...</td>\n",
       "      <td>4.77</td>\n",
       "      <td>4.27</td>\n",
       "      <td>3.98</td>\n",
       "      <td>2</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>3.62</td>\n",
       "      <td>3.78</td>\n",
       "      <td></td>\n",
       "      <td>3.52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DefenCorps</th>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td></td>\n",
       "      <td>4.07</td>\n",
       "      <td>4.37</td>\n",
       "      <td>4.32</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>2.14</td>\n",
       "      <td>...</td>\n",
       "      <td>4.57</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.13</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Derek</th>\n",
       "      <td></td>\n",
       "      <td>4.35</td>\n",
       "      <td>4.43</td>\n",
       "      <td>4.18</td>\n",
       "      <td>4.42</td>\n",
       "      <td>4.45</td>\n",
       "      <td>4.42</td>\n",
       "      <td>4.15</td>\n",
       "      <td>3.11</td>\n",
       "      <td>3.2</td>\n",
       "      <td>...</td>\n",
       "      <td>3.45</td>\n",
       "      <td></td>\n",
       "      <td>4.1</td>\n",
       "      <td></td>\n",
       "      <td>4</td>\n",
       "      <td>2.44</td>\n",
       "      <td>4.21</td>\n",
       "      <td>3.98</td>\n",
       "      <td></td>\n",
       "      <td>4.22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dope</th>\n",
       "      <td></td>\n",
       "      <td>4.1</td>\n",
       "      <td></td>\n",
       "      <td>4.07</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.38</td>\n",
       "      <td>4.35</td>\n",
       "      <td></td>\n",
       "      <td>3.75</td>\n",
       "      <td></td>\n",
       "      <td>...</td>\n",
       "      <td>4.38</td>\n",
       "      <td>3.73</td>\n",
       "      <td>4.1</td>\n",
       "      <td>2.52</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>50 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 rating                                                        \\\n",
       "beer                6     30    31    33    34    39    59    61    63    65    \n",
       "user                                                                            \n",
       "57md               3.69                    3.63                       4  2.18   \n",
       "ALC82                                       4.5  4.73               3.9  2.25   \n",
       "AgentMunky         4.07  3.95  3.16  3.88   4.3        4.15        3.46  1.72   \n",
       "Alieniloquium            4.07        3.88  4.12  3.95  3.41                     \n",
       "Arbitrator               2.51  3.27  3.56  4.28  4.32  4.01     4  4.17         \n",
       "ArrogantB                                        4.57   4.1  3.36               \n",
       "BARFLYB                  3.77        3.72  4.35         4.6     4        2.76   \n",
       "BB1313              3.6                          4.05        3.64  3.58         \n",
       "BEER88                      4        4.25  4.25  4.25   4.5  4.13   3.5     2   \n",
       "BEERMILER12              4.42   3.9  4.19   4.6              4.35  4.13  2.95   \n",
       "BEERchitect           3  4.08  3.93   4.2  3.95  4.21  3.83  4.15  3.53  2.18   \n",
       "BMMillsy                 4.31  3.73         4.2  4.27  3.86         3.1         \n",
       "BarryMFBurton      3.95  4.02        4.47  4.15  4.09        4.24  3.61  1.92   \n",
       "BeerFMAndy         3.86  4.74  3.97  4.34  4.53  4.15  4.09  4.26  4.21  1.78   \n",
       "BeerLover99              4.32  3.83   4.5  4.95  4.42   4.2        3.98  2.35   \n",
       "Beerandraiderfan   2.63  3.44   2.5  3.23  2.38  4.27  3.06     4  3.73   2.4   \n",
       "Beginner2          3.86   4.1  3.76  4.13  4.38  4.29  4.31  4.18  3.95  2.72   \n",
       "Bierguy5                  3.9                    4.25              3.78         \n",
       "BigMike                                       4                    3.64   1.4   \n",
       "Biggiekrock627                          4  4.49  3.97  3.68        4.16   1.7   \n",
       "Bighuge               3  4.59  3.52  4.22  4.33  3.15  4.52  3.11  4.38   1.5   \n",
       "Billolick          3.41  4.02   4.3  4.19  4.08   4.2   4.1  4.15  3.98  2.28   \n",
       "Bitterbill                3.9        4.23  4.22  4.65  4.45  4.47  4.13  3.38   \n",
       "Bouleboubier       4.12  4.17  3.95  4.54  4.92        3.88  4.85  3.93         \n",
       "Brad007            3.41  4.13  4.85  3.93  4.82        3.88  2.81  3.12   2.1   \n",
       "Brenden             3.6   4.4   4.1  4.37  4.73  4.47              4.03  2.73   \n",
       "Brent              3.06  4.06     4   4.6     4  4.24         4.8     4   1.1   \n",
       "BretSikkink              3.88  4.05   4.6  4.85  4.38  4.18  4.23  3.63  2.97   \n",
       "BucannonXC5        3.76  4.16        4.01  4.27        4.05  3.76  3.61  2.54   \n",
       "Buck89                   4.34              4.35  4.24  4.28  3.94  3.88  2.29   \n",
       "BuckeyeNation      2.93  3.92  3.42   4.2  4.05   4.6  4.57  4.35  3.65  1.62   \n",
       "CBrodbeck                4.13        3.94  4.18  4.13        4.04  3.77         \n",
       "CHickman           3.67   4.2              4.35  4.45        4.48  3.81  2.74   \n",
       "ChainGangGuy       3.23  4.29  3.53  3.78  4.02        3.25  3.68  3.53         \n",
       "Chaney                   4.32     4  4.25  4.32  4.85  4.38                     \n",
       "CharlesK              3     4        3.94  4.14  4.11     4  3.84               \n",
       "Chico1985                3.98        4.33  4.33  4.03  4.45        3.75  1.63   \n",
       "ColForbinBC              4.32  3.92  4.18  4.18   4.4   4.2           4  1.88   \n",
       "Cpmitchno1                                 4.42     4  4.17                     \n",
       "CrellMoset          3.5  4.25               4.4   4.5  4.35                     \n",
       "CuriousMonk           3  4.05         4.5   3.7  4.47     4     4   3.8         \n",
       "Cyberkedi             4  4.41   4.2   4.2  4.02        4.08  4.41  3.54  2.17   \n",
       "DIM                      4.37     4   3.8  4.25  4.85                           \n",
       "Damian                    4.3        4.24  4.69  4.65  3.66  4.17  3.58  2.81   \n",
       "Daniellobo         3.38  4.07           4  3.28        3.98   4.1  3.62         \n",
       "Darkmagus82        3.85  4.14  3.75  4.09  4.37  4.03  3.89   3.9  3.87  2.85   \n",
       "DavoleBomb                3.7        4.33  4.52   4.2  4.39              1.89   \n",
       "DefenCorps                  4        4.07  4.37  4.32                    2.14   \n",
       "Derek                    4.35  4.43  4.18  4.42  4.45  4.42  4.15  3.11   3.2   \n",
       "Dope                      4.1        4.07   4.1  4.38  4.35        3.75         \n",
       "\n",
       "                  ...                                                         \\\n",
       "beer              ...    226   228   229   232   236   246   248   276   279   \n",
       "user              ...                                                          \n",
       "57md              ...   3.87  3.79  4.29  1.69  3.61              3.97  4.06   \n",
       "ALC82             ...   4.28                 3        3.75        4.06         \n",
       "AgentMunky        ...   3.85  4.43  3.68  1.96  3.33               4.4         \n",
       "Alieniloquium     ...    4.2                                                   \n",
       "Arbitrator        ...   4.23  4.48   4.5                     3.7        3.41   \n",
       "ArrogantB         ...   3.36  3.88                                4.43   4.3   \n",
       "BARFLYB           ...   4.08  3.85              3.43         3.5  3.92  2.93   \n",
       "BB1313            ...   4.42  4.26     4                          3.81         \n",
       "BEER88            ...   4.25  3.96         2.5        2.46  3.21  3.75  3.53   \n",
       "BEERMILER12       ...                     2.38        2.91        4.13     4   \n",
       "BEERchitect       ...      4  4.35  4.28  1.76     4  2.98  4.45  4.42   4.1   \n",
       "BMMillsy          ...   4.36  4.27  4.26        3.82        3.59  3.71  3.51   \n",
       "BarryMFBurton     ...         3.85  3.72        3.77  2.56        3.95  3.77   \n",
       "BeerFMAndy        ...   4.78  4.89  4.36  2.36        3.26   3.8  4.63  4.35   \n",
       "BeerLover99       ...   4.23  4.43  3.91     2              3.92  4.42  3.63   \n",
       "Beerandraiderfan  ...   4.22  3.43  3.58  1.57  3.28  2.97  3.82  4.83  4.22   \n",
       "Beginner2         ...   4.12     4  4.16  2.25  3.99  3.13     4  3.88  3.75   \n",
       "Bierguy5          ...   4.52  4.13  4.38   2.4        2.35           4         \n",
       "BigMike           ...    3.6              1.53        2.25           4  3.59   \n",
       "Biggiekrock627    ...   3.98  4.13  3.69  1.08        2.38  3.65  3.77         \n",
       "Bighuge           ...    4.5     5  4.53                    4.28  4.47  4.33   \n",
       "Billolick         ...                     2.68  3.87   2.9   4.1  4.27  4.25   \n",
       "Bitterbill        ...   4.68              1.31  4.32  3.45   3.2  4.18  4.13   \n",
       "Bouleboubier      ...    4.5  4.15  4.32  2.87  3.89  3.92   4.3  4.15   3.9   \n",
       "Brad007           ...         3.88  4.38     1  3.55   3.7  4.85   3.2  3.73   \n",
       "Brenden           ...    4.2   4.3   4.2  1.52  3.68  2.85  3.63  4.12  3.98   \n",
       "Brent             ...      4     4     4  1.84        1.82   4.4   4.4         \n",
       "BretSikkink       ...                     2.35  3.43  2.72  3.85   4.1   3.9   \n",
       "BucannonXC5       ...                     2.56        1.68   3.9  4.18   3.8   \n",
       "Buck89            ...                     1.85                    4.16         \n",
       "BuckeyeNation     ...   4.83  4.57  4.65  1.88  3.33     3  3.97     4   4.1   \n",
       "CBrodbeck         ...          4.1        2.31        2.47  3.71   4.1         \n",
       "CHickman          ...                     1.81        3.17  3.98  4.67  4.39   \n",
       "ChainGangGuy      ...   4.15  3.93  4.08  1.96  3.18  2.68   3.3  3.52  3.03   \n",
       "Chaney            ...   4.48  4.33   4.2                                       \n",
       "CharlesK          ...   3.11  2.58   3.5  2.47        1.64  2.84  3.44  3.29   \n",
       "Chico1985         ...   4.35              2.74         2.2  4.03  4.13         \n",
       "ColForbinBC       ...   4.32   4.1                                3.98         \n",
       "Cpmitchno1        ...   4.12        4.02                          3.96         \n",
       "CrellMoset        ...   4.42  4.08     4              3.01        3.92  3.83   \n",
       "CuriousMonk       ...   3.82  3.65     4  2.93         2.6        3.75  3.75   \n",
       "Cyberkedi         ...   4.33  3.72        2.67  3.91  3.05        4.06  3.83   \n",
       "DIM               ...   4.47                                   4           4   \n",
       "Damian            ...    3.6  3.56        2.89        1.94  4.08  2.96         \n",
       "Daniellobo        ...         4.21  4.13                          3.98  3.83   \n",
       "Darkmagus82       ...   4.34  3.83  4.11        3.93  2.21   3.7  3.96  3.77   \n",
       "DavoleBomb        ...   4.77  4.27  3.98     2              3.62  3.78         \n",
       "DefenCorps        ...   4.57  4.25  4.13                                       \n",
       "Derek             ...   3.45         4.1           4  2.44  4.21  3.98         \n",
       "Dope              ...   4.38  3.73   4.1  2.52                                 \n",
       "\n",
       "                        \n",
       "beer               283  \n",
       "user                    \n",
       "57md              4.13  \n",
       "ALC82                   \n",
       "AgentMunky        3.26  \n",
       "Alieniloquium           \n",
       "Arbitrator              \n",
       "ArrogantB            4  \n",
       "BARFLYB           3.78  \n",
       "BB1313            3.61  \n",
       "BEER88            4.32  \n",
       "BEERMILER12       4.22  \n",
       "BEERchitect       4.33  \n",
       "BMMillsy                \n",
       "BarryMFBurton     4.27  \n",
       "BeerFMAndy        4.84  \n",
       "BeerLover99             \n",
       "Beerandraiderfan  4.41  \n",
       "Beginner2            4  \n",
       "Bierguy5                \n",
       "BigMike                 \n",
       "Biggiekrock627    3.79  \n",
       "Bighuge           4.15  \n",
       "Billolick         4.25  \n",
       "Bitterbill        4.33  \n",
       "Bouleboubier      3.96  \n",
       "Brad007           3.68  \n",
       "Brenden                 \n",
       "Brent              4.9  \n",
       "BretSikkink       4.08  \n",
       "BucannonXC5             \n",
       "Buck89            4.34  \n",
       "BuckeyeNation      4.1  \n",
       "CBrodbeck          4.1  \n",
       "CHickman           4.6  \n",
       "ChainGangGuy      3.06  \n",
       "Chaney                  \n",
       "CharlesK                \n",
       "Chico1985         3.88  \n",
       "ColForbinBC             \n",
       "Cpmitchno1        4.08  \n",
       "CrellMoset        4.08  \n",
       "CuriousMonk          4  \n",
       "Cyberkedi         4.27  \n",
       "DIM               3.76  \n",
       "Damian            2.52  \n",
       "Daniellobo        3.88  \n",
       "Darkmagus82       4.12  \n",
       "DavoleBomb        3.52  \n",
       "DefenCorps              \n",
       "Derek             4.22  \n",
       "Dope                    \n",
       "\n",
       "[50 rows x 50 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_table.iloc[:50,:50].fillna(\" \")\n",
    "#이러한 평점행렬 R은 데이터가 일부 위치에만 존재하는 희소행렬(sparse matrix)\n",
    "#(50x50) 개만 뽑음."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#시각화\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "plt.imshow(df_table)\n",
    "plt.grid(False)\n",
    "plt.xlabel(\"item\")\n",
    "plt.ylabel(\"user\")\n",
    "plt.title(\"Rate Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 모델 제작 순서\n",
    "- 데이터셋의 split, folds 메서드를 사용해 k-fold 트레이닝 데이터셋과 테스트 데이터셋 만듦\n",
    "- 모형 알고리즘 객체 생성\n",
    "- training 데이터셋으로 모수를 추정한 후, test 메서드로 테스트 데이터셋에 대한 예측 실시\n",
    "- accuracy 계산\n",
    "\n",
    "- ★(evalueate 로 한번에도 가능)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting surprise\n",
      "  Using cached https://files.pythonhosted.org/packages/61/de/e5cba8682201fcf9c3719a6fdda95693468ed061945493dea2dd37c5618b/surprise-0.1-py2.py3-none-any.whl\n",
      "Requirement already satisfied: scikit-surprise in c:\\programdata\\anaconda3\\lib\\site-packages (from surprise) (1.0.6)\n",
      "Requirement already satisfied: joblib>=0.11 in c:\\programdata\\anaconda3\\lib\\site-packages (from scikit-surprise->surprise) (0.13.2)\n",
      "Requirement already satisfied: numpy>=1.11.2 in c:\\programdata\\anaconda3\\lib\\site-packages (from scikit-surprise->surprise) (1.15.4)\n",
      "Requirement already satisfied: scipy>=1.0.0 in c:\\programdata\\anaconda3\\lib\\site-packages (from scikit-surprise->surprise) (1.1.0)\n",
      "Requirement already satisfied: six>=1.10.0 in c:\\users\\suyn3\\appdata\\roaming\\python\\python37\\site-packages (from scikit-surprise->surprise) (1.12.0)\n",
      "Installing collected packages: surprise\n",
      "Successfully installed surprise-0.1\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: You are using pip version 19.2.3, however version 19.3.1 is available.\n",
      "You should consider upgrading via the 'python -m pip install --upgrade pip' command.\n"
     ]
    }
   ],
   "source": [
    "!pip install surprise\n",
    "import surprise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import Reader, Dataset\n",
    "from surprise.model_selection import train_test_split\n",
    "\n",
    "reader = Reader(line_format='user item rating', sep=',', rating_scale=(0.5,5))\n",
    "#데이터의 칼럼 순서는 uid, iid, rating 순서를 반드시 지켜야 함\n",
    "data = Dataset.load_from_df(df[['userid','beer','rating']],reader)\n",
    "\n",
    "trainset, testset = train_test_split(data, test_size=.25, random_state=0) #train, test set 분할"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 베이스라인 모형"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise.model_selection import KFold\n",
    "from surprise import BaselineOnly\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Estimating biases using als...\n",
      "RMSE: 0.3396\n",
      "Estimating biases using als...\n",
      "RMSE: 0.3333\n",
      "Estimating biases using als...\n",
      "RMSE: 0.3370\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.3366439478906325"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "bsl_options = {\n",
    "    'method': 'als',\n",
    "    'n_epochs': 5,\n",
    "    'reg_u': 12,\n",
    "    'reg_i': 5\n",
    "}\n",
    "algo = surprise.BaselineOnly(bsl_options)\n",
    "\n",
    "np.random.seed(0)\n",
    "acc = np.zeros(3)\n",
    "cv = KFold(3)\n",
    "for i, (trainset, testset) in enumerate(cv.split(data)):\n",
    "    algo.fit(trainset)\n",
    "    predictions = algo.test(testset)\n",
    "    acc[i] = surprise.accuracy.rmse(predictions, verbose=True)\n",
    "acc.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Estimating biases using als...\n",
      "Estimating biases using als...\n",
      "Estimating biases using als...\n",
      "Estimating biases using als...\n",
      "Estimating biases using als...\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.33391078, 0.33564717, 0.33871842, 0.33571915, 0.33760788]),\n",
       " 'test_mae': array([0.24006561, 0.23816496, 0.24181065, 0.24070886, 0.2412288 ]),\n",
       " 'fit_time': (0.16755318641662598,\n",
       "  0.18550419807434082,\n",
       "  0.20844244956970215,\n",
       "  0.1506333351135254,\n",
       "  0.1486036777496338),\n",
       " 'test_time': (0.3390920162200928,\n",
       "  0.41887831687927246,\n",
       "  0.3191823959350586,\n",
       "  0.21837782859802246,\n",
       "  0.16552114486694336)}"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#위의 코드 한번에\n",
    "from surprise.model_selection import cross_validate\n",
    "cross_validate(algo, data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Neighborhood 모형\n",
    "<옵션>\n",
    "- name: 사용할 유사도의 종류, 디폴트는 'MSD'\n",
    "- user_based : True면 사용자 기반, False면 상품 기반\n",
    "- min_support : 두 사용자나 상품에서 공통적으로 있는 평점 원소 수의 최솟값.\n",
    "- shrinkage : shrinkage 가중치. 디폴트는 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import KNNBasic\n",
    "from surprise.model_selection import cross_validate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.36395562, 0.36613159, 0.3626647 , 0.36377216, 0.36405512]),\n",
       " 'test_mae': array([0.25703913, 0.25690976, 0.25611493, 0.25782994, 0.25679896]),\n",
       " 'fit_time': (2.9042389392852783,\n",
       "  2.873318910598755,\n",
       "  2.8663389682769775,\n",
       "  3.1096932888031006,\n",
       "  2.9949615001678467),\n",
       " 'test_time': (8.325753688812256,\n",
       "  9.339260578155518,\n",
       "  8.657899618148804,\n",
       "  9.058826923370361,\n",
       "  8.02356219291687)}"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#우리는 코사인 유사도 사용!\n",
    "sim_options = {'name': 'cosine'}\n",
    "algo = surprise.KNNBasic(sim_options=sim_options)\n",
    "cross_validate(algo, data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## KNN 가중치 예측 방법\n",
    "- KNNBasic : 평점들을 단순히 가중 평균함, N^k는 k개의 가장 유사도가 큰 벡터의 집합\n",
    "- KNNWithMeans : 평점들을 평균값 기준으로 가중 평균함\n",
    "- KNNBaseline : 평점들을 베이스라인 모형의 값 기준으로 가중 평균함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import KNNWithMeans\n",
    "from surprise import KNNBaseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.34245794, 0.33727805, 0.33479002, 0.34338102, 0.33852872]),\n",
       " 'test_mae': array([0.24605065, 0.2446332 , 0.24367377, 0.24841396, 0.24476033]),\n",
       " 'fit_time': (3.117666244506836,\n",
       "  2.927178382873535,\n",
       "  2.852341413497925,\n",
       "  2.807464122772217,\n",
       "  3.004972457885742),\n",
       " 'test_time': (9.637250900268555,\n",
       "  9.168533563613892,\n",
       "  8.135295152664185,\n",
       "  9.05982756614685,\n",
       "  8.93811559677124)}"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sim_options = {'name': 'cosine'}\n",
    "algo = surprise.KNNWithMeans(sim_options=sim_options)\n",
    "cross_validate(algo, data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Estimating biases using als...\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Estimating biases using als...\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Estimating biases using als...\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Estimating biases using als...\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n",
      "Estimating biases using als...\n",
      "Computing the cosine similarity matrix...\n",
      "Done computing similarity matrix.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.33012503, 0.33180334, 0.33894008, 0.33964316, 0.3331826 ]),\n",
       " 'test_mae': array([0.23667383, 0.23764794, 0.2416309 , 0.2418265 , 0.2374919 ]),\n",
       " 'fit_time': (3.393930196762085,\n",
       "  3.3889429569244385,\n",
       "  4.392226696014404,\n",
       "  2.8603618144989014,\n",
       "  2.9062671661376953),\n",
       " 'test_time': (10.941730499267578,\n",
       "  10.488970041275024,\n",
       "  13.005803108215332,\n",
       "  9.574503421783447,\n",
       "  9.968329429626465)}"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sim_options = {'name': 'cosine'}\n",
    "algo = surprise.KNNBaseline(sim_options=sim_options)\n",
    "cross_validate(algo, data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 잠재 요인 방법(Latent Factor)\n",
    "- 사용자의 특성 벡터나 상품의 특성 벡터의 길이는 수천에서 수십억에 달하는 긴 크기가 될 수도 있다.\n",
    "- Latent Factor 모형은 이렇게 긴 사용자 특성이나 상품 특성을 몇 개의 요인 벡터로 간략화(approximate)할 수 있다는 가정에서 출발한 모형이다.\n",
    "- PCA(Principle Component Analysis)를 사용하면 긴 특성 벡터를 소수의 차원으로 차원 축소할 수 있듯이 사용자의 특성도 차원 축소 할 수 있다.\n",
    "\n",
    "- 영화에 대한 평점을 주는 경우, 코미디, 액션, 드라마 등 몇개의 장르 요인이 있어서 사용자는 특정한 장르 요소에 대해 더 점수를 많이 주거나 적게 줄 수 있다. 그리고 영화 자체도 이러한 장르 요인을 가지고 있다면 해당 사용자의 그 영화에 대한 평점은 사용자의 장르 요인 벡터와 영화의 장르 요인 벡터의 내적으로 표시할 수 있다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. NMF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.73958185, 0.7442849 , 0.74048016, 0.74314277, 0.7321391 ]),\n",
       " 'test_mae': array([0.65931511, 0.66244435, 0.65857706, 0.66147633, 0.65264787]),\n",
       " 'fit_time': (16.08202576637268,\n",
       "  15.507524728775024,\n",
       "  16.544325351715088,\n",
       "  19.686875343322754,\n",
       "  19.193224668502808),\n",
       " 'test_time': (0.2034592628479004,\n",
       "  0.24936938285827637,\n",
       "  0.22141075134277344,\n",
       "  0.44281578063964844,\n",
       "  0.5305829048156738)}"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from surprise import NMF\n",
    "algo = surprise.NMF(n_factors=50)\n",
    "cross_validate(algo, data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7389788234720699"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "algo = surprise.NMF(n_factors=50)\n",
    "cross_validate(algo, data)['test_rmse'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. SVD(Singular Value Decomposition)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import Reader, Dataset\n",
    "\n",
    "reader = Reader(line_format='user item rating', sep=',', rating_scale=(0.5,5))\n",
    "#데이터의 칼럼 순서는 uid, iid, rating 순서를 반드시 지켜야 함\n",
    "data = Dataset.load_from_df(df[['int_user','beer','rating']],reader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainset, testset = train_test_split(data, test_size=.25, random_state=0) #train, test set 분할"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise import SVD\n",
    "from surprise import accuracy\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SVD 클래스의 입력 파라미터\n",
    "- n_factors: 잠재요인 k의 개수, 디폴트: 100, 커질수록 정확도가 높아질 수 있으나 과적합 문제 발생 가능\n",
    "- n_epochs: SGD, 디폴트: 20\n",
    "- biased : 베이스라인 사용자 편향 적용 여부, 디폴트: True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.34233825, 0.33834952, 0.33394751, 0.33504995, 0.3382407 ]),\n",
       " 'test_mae': array([0.24322278, 0.24272906, 0.23993683, 0.24038491, 0.24241232]),\n",
       " 'fit_time': (9.126651763916016,\n",
       "  5.745615720748901,\n",
       "  5.960139751434326,\n",
       "  5.223080635070801,\n",
       "  5.933145999908447),\n",
       " 'test_time': (0.3291149139404297,\n",
       "  0.3400897979736328,\n",
       "  0.20545029640197754,\n",
       "  0.27127552032470703,\n",
       "  0.20943760871887207)}"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "algo = surprise.SVD(n_factors=50)\n",
    "cross_validate(algo, data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.3374783596637191"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "algo = surprise.SVD(n_factors=50)\n",
    "cross_validate(algo, data)['test_rmse'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE: 0.3371\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.33705645097044357"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#SVD를 이용해 학습 수행\n",
    "trainset, testset = train_test_split(data, test_size=.25, random_state=0) #train, test set 분할\n",
    "\n",
    "algo = SVD(n_factors=50, random_state=0)\n",
    "algo.fit(trainset)\n",
    "predictions = algo.test(testset)\n",
    "accuracy.rmse(predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 제대로 모델이 만들어졌을까? 한번 추천해보자."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "#유저가 평가하지 않은 beer목록을 반환해 주는 함수 : get_unbeer_surprise\n",
    "\n",
    "def get_unbeer_surprise(df, userid):\n",
    "    beers = df[df['userid']==userid]['beer'].unique().tolist() #해당 유저가 평가한 beer\n",
    "    \n",
    "    total_beers = df['beer'].unique().tolist() #모든 beer\n",
    "    \n",
    "    unbeers = [beer for beer in total_beers if beer not in beers] #해당 유저가 평가안한 beer\n",
    "    \n",
    "    print('이미 평가한 beer 수: ', len(beers), '안 평가한 beer 수: ' , len(unbeers), '전체 beer 수: ', len(total_beers))\n",
    "    \n",
    "    return unbeers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "이미 평가한 beer 수:  317 안 평가한 beer 수:  164 전체 beer 수:  481\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[125646,\n",
       " 58610,\n",
       " 50697,\n",
       " 117177,\n",
       " 100217,\n",
       " 88969,\n",
       " 195,\n",
       " 86237,\n",
       " 1493,\n",
       " 2264,\n",
       " 132,\n",
       " 142,\n",
       " 28577,\n",
       " 449,\n",
       " 228,\n",
       " 232,\n",
       " 276,\n",
       " 162643,\n",
       " 332,\n",
       " 411,\n",
       " 429,\n",
       " 1798,\n",
       " 78011,\n",
       " 582,\n",
       " 593,\n",
       " 916,\n",
       " 639,\n",
       " 650,\n",
       " 100421,\n",
       " 712,\n",
       " 751,\n",
       " 808,\n",
       " 175577,\n",
       " 811,\n",
       " 115317,\n",
       " 1003,\n",
       " 1963,\n",
       " 72170,\n",
       " 1256,\n",
       " 45897,\n",
       " 176650,\n",
       " 1421,\n",
       " 113523,\n",
       " 1577,\n",
       " 1585,\n",
       " 187317,\n",
       " 38180,\n",
       " 32286,\n",
       " 2729,\n",
       " 144200,\n",
       " 48385,\n",
       " 148052,\n",
       " 41815,\n",
       " 7597,\n",
       " 2137,\n",
       " 3558,\n",
       " 33127,\n",
       " 48139,\n",
       " 202078,\n",
       " 98383,\n",
       " 146788,\n",
       " 64550,\n",
       " 11922,\n",
       " 30282,\n",
       " 16062,\n",
       " 19589,\n",
       " 138735,\n",
       " 20781,\n",
       " 21950,\n",
       " 46363,\n",
       " 24905,\n",
       " 41121,\n",
       " 48582,\n",
       " 76393,\n",
       " 27604,\n",
       " 197183,\n",
       " 84045,\n",
       " 133621,\n",
       " 68269,\n",
       " 63249,\n",
       " 45653,\n",
       " 178857,\n",
       " 150209,\n",
       " 59,\n",
       " 99,\n",
       " 179482,\n",
       " 47022,\n",
       " 34688,\n",
       " 225,\n",
       " 79286,\n",
       " 114804,\n",
       " 611,\n",
       " 615,\n",
       " 45973,\n",
       " 656,\n",
       " 672,\n",
       " 837,\n",
       " 46849,\n",
       " 16909,\n",
       " 154542,\n",
       " 111616,\n",
       " 27804,\n",
       " 17060,\n",
       " 127652,\n",
       " 4073,\n",
       " 5441,\n",
       " 57747,\n",
       " 7971,\n",
       " 150877,\n",
       " 182256,\n",
       " 9474,\n",
       " 12770,\n",
       " 16115,\n",
       " 16814,\n",
       " 61062,\n",
       " 57908,\n",
       " 17900,\n",
       " 211516,\n",
       " 88949,\n",
       " 29209,\n",
       " 22227,\n",
       " 22790,\n",
       " 62761,\n",
       " 155168,\n",
       " 35738,\n",
       " 55862,\n",
       " 75446,\n",
       " 107406,\n",
       " 5653,\n",
       " 205,\n",
       " 246,\n",
       " 248,\n",
       " 311,\n",
       " 47434,\n",
       " 6108,\n",
       " 1054,\n",
       " 152542,\n",
       " 1346,\n",
       " 1361,\n",
       " 1445,\n",
       " 3523,\n",
       " 73764,\n",
       " 35626,\n",
       " 2178,\n",
       " 935,\n",
       " 1836,\n",
       " 77708,\n",
       " 131782,\n",
       " 102340,\n",
       " 110895,\n",
       " 5281,\n",
       " 47731,\n",
       " 10386,\n",
       " 139521,\n",
       " 60426,\n",
       " 59173,\n",
       " 100443,\n",
       " 149331,\n",
       " 92960,\n",
       " 123062,\n",
       " 134064,\n",
       " 113674,\n",
       " 116067,\n",
       " 145496]"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_unbeer_surprise(df, 190)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['id', 'name', 'brewery_id', 'state', 'country', 'style', 'availability',\n",
       "       'abv', 'notes', 'retired'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "beers = pd.read_csv('beers.csv')\n",
    "beers.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>name</th>\n",
       "      <th>brewery_id</th>\n",
       "      <th>state</th>\n",
       "      <th>country</th>\n",
       "      <th>style</th>\n",
       "      <th>availability</th>\n",
       "      <th>abv</th>\n",
       "      <th>notes</th>\n",
       "      <th>retired</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>202522</td>\n",
       "      <td>Olde Cogitator</td>\n",
       "      <td>2199</td>\n",
       "      <td>CA</td>\n",
       "      <td>US</td>\n",
       "      <td>English Oatmeal Stout</td>\n",
       "      <td>Rotating</td>\n",
       "      <td>7.3</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>f</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>82352</td>\n",
       "      <td>Konrads Stout Russian Imperial Stout</td>\n",
       "      <td>18604</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NO</td>\n",
       "      <td>Russian Imperial Stout</td>\n",
       "      <td>Rotating</td>\n",
       "      <td>10.4</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>f</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>214879</td>\n",
       "      <td>Scottish Right</td>\n",
       "      <td>44306</td>\n",
       "      <td>IN</td>\n",
       "      <td>US</td>\n",
       "      <td>Scottish Ale</td>\n",
       "      <td>Year-round</td>\n",
       "      <td>4.0</td>\n",
       "      <td>No notes at this time.</td>\n",
       "      <td>t</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       id                                  name  brewery_id state country  \\\n",
       "0  202522                        Olde Cogitator        2199    CA      US   \n",
       "1   82352  Konrads Stout Russian Imperial Stout       18604   NaN      NO   \n",
       "2  214879                        Scottish Right       44306    IN      US   \n",
       "\n",
       "                    style availability   abv                   notes retired  \n",
       "0   English Oatmeal Stout     Rotating   7.3  No notes at this time.       f  \n",
       "1  Russian Imperial Stout     Rotating  10.4  No notes at this time.       f  \n",
       "2            Scottish Ale   Year-round   4.0  No notes at this time.       t  "
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "beers.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "def recomm_beer_by_surprise(algo, userid, unbeers, top_n=10):\n",
    "    \n",
    "    predictions = [algo.predict(userid, beer) for beer in unbeers] #대입된 userid가 평가 안한 beer에 algo.predict 반복 적용\n",
    "    \n",
    "    def sortkey_est(pred):\n",
    "        return pred.est #앞의 prediciton에서 예상 평점인 pred.est만 반환\n",
    "    \n",
    "    predictions.sort(key=sortkey_est, reverse=True) #predicions를 pred.est값에 따라 sorting함. reverse=True이므로 내림차순 정렬(평점 큰 순서로)\n",
    "    top_predictions=predictions[:top_n] #top_n으로 대입된 숫자까지만 predictions를 자름\n",
    "    \n",
    "    #top_n으로 추출된 영화아이디, 예상 평점, 제목 추출\n",
    "    top_beer_ids = [ int(pred.iid) for pred in top_predictions ]\n",
    "    top_beer_rating = [ pred.est for pred in top_predictions ]\n",
    "    top_beer_titles = beers[beers.id.isin(top_beer_ids)]['name'] #데이터셋 필요\n",
    "    top_beer_country = beers[beers.id.isin(top_beer_ids)]['country']\n",
    "    top_beer_preds = [ (id, title, rating, country) for id, title, rating, country in zip(top_beer_ids, top_beer_titles, top_beer_rating, top_beer_country) ] \n",
    "    #앞에 정의한 것들을 id, 제목, 평점, 국가 순으로 반환\n",
    "    \n",
    "    return top_beer_preds\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "이미 평가한 beer 수:  317 안 평가한 beer 수:  164 전체 beer 수:  481\n"
     ]
    }
   ],
   "source": [
    "unbeers = get_unbeer_surprise(df, 190)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_beer_preds = recomm_beer_by_surprise(algo, 190, unbeers, top_n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(115317, 'Julius', 4.723246499950317, 'US'),\n",
       " (131782, \"Cantillon Fou' Foune\", 4.692933420227621, 'BE'),\n",
       " (16814, 'Dinner', 4.67549470229658, 'US'),\n",
       " (57747, 'Parabola', 4.67191888617864, 'US'),\n",
       " (5281, 'Bourbon County Brand Coffee Stout', 4.662119510318055, 'US'),\n",
       " (86237, 'Pliny The Elder', 4.655880983447909, 'US'),\n",
       " (134064, 'Heady Topper', 4.653543163766593, 'US'),\n",
       " (7971, 'Everett Porter', 4.638285264603047, 'US'),\n",
       " (61062, 'Double Dry Hopped Congress Street', 4.627296484268649, 'US'),\n",
       " (41815, 'Vanilla Rye Bourbon County Brand Stout', 4.620966132576913, 'US')]"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_beer_preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<<TOP-10 추천 맥주 리스트>>\n",
      "추천맥주:  Julius  / 예상 평점은:  4.723  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Cantillon Fou' Foune  / 예상 평점은:  4.693  / 이 맥주의 판매국:  BE\n",
      "추천맥주:  Dinner  / 예상 평점은:  4.675  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Parabola  / 예상 평점은:  4.672  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Bourbon County Brand Coffee Stout  / 예상 평점은:  4.662  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Pliny The Elder  / 예상 평점은:  4.656  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Heady Topper  / 예상 평점은:  4.654  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Everett Porter  / 예상 평점은:  4.638  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Double Dry Hopped Congress Street  / 예상 평점은:  4.627  / 이 맥주의 판매국:  US\n",
      "추천맥주:  Vanilla Rye Bourbon County Brand Stout  / 예상 평점은:  4.621  / 이 맥주의 판매국:  US\n"
     ]
    }
   ],
   "source": [
    "print('<<TOP-10 추천 맥주 리스트>>')\n",
    "for top_beer in top_beer_preds:\n",
    "    print(\"추천맥주: \", top_beer[1],\" / 예상 평점은: \", np.round(top_beer[2],3), \" / 이 맥주의 판매국: \", top_beer[3])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 교차 검증과 하이퍼 파라미터 튜닝(모델 성능 평가) : surprise.model_selection\n",
    "- cross_validate() : 교차 검증\n",
    "- GridSearchCV() : 하이퍼 파라미터 튜닝(n_epochs, n_factors 지정)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "from surprise.model_selection import cross_validate\n",
    "from surprise.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Evaluating RMSE, MAE of algorithm SVD on 10 split(s).\n",
      "\n",
      "                  Fold 1  Fold 2  Fold 3  Fold 4  Fold 5  Fold 6  Fold 7  Fold 8  Fold 9  Fold 10 Mean    Std     \n",
      "RMSE (testset)    0.3306  0.3399  0.3409  0.3381  0.3354  0.3304  0.3389  0.3353  0.3404  0.3398  0.3370  0.0037  \n",
      "MAE (testset)     0.2401  0.2436  0.2434  0.2408  0.2390  0.2384  0.2427  0.2395  0.2430  0.2417  0.2412  0.0018  \n",
      "Fit time          5.74    5.66    6.18    7.08    8.89    7.36    6.98    7.32    7.60    7.04    6.99    0.91    \n",
      "Test time         0.12    0.13    0.13    0.46    0.18    0.11    0.23    0.21    0.12    0.13    0.18    0.10    \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'test_rmse': array([0.33061244, 0.33992956, 0.34091473, 0.33807869, 0.33538209,\n",
       "        0.33043921, 0.33889182, 0.33531074, 0.34044412, 0.33978929]),\n",
       " 'test_mae': array([0.24014794, 0.24363188, 0.24339404, 0.24084642, 0.23896127,\n",
       "        0.238383  , 0.2426549 , 0.23948309, 0.24299924, 0.24173532]),\n",
       " 'fit_time': (5.739627838134766,\n",
       "  5.66485595703125,\n",
       "  6.181442975997925,\n",
       "  7.081078052520752,\n",
       "  8.893236637115479,\n",
       "  7.356345176696777,\n",
       "  6.97639274597168,\n",
       "  7.324396371841431,\n",
       "  7.602684020996094,\n",
       "  7.043180227279663),\n",
       " 'test_time': (0.1177225112915039,\n",
       "  0.12865924835205078,\n",
       "  0.13267803192138672,\n",
       "  0.46475768089294434,\n",
       "  0.17752790451049805,\n",
       "  0.11466121673583984,\n",
       "  0.2293529510498047,\n",
       "  0.20544981956481934,\n",
       "  0.11569094657897949,\n",
       "  0.13164973258972168)}"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#10 fold CV\n",
    "cross_validate(algo, data, measures = ['RMSE','MAE'], cv=10, verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "#GridSearch \n",
    "param_grid = {'n_epochs':[20,40,60], 'n_factors':[20,50,100]} #최적화할 파라미터 딕셔너리 형태로 지정\n",
    "gs = GridSearchCV(SVD, param_grid, measures = ['rmse'], cv=10) #CV를 10개 폴트 세트로 지정, 성능 평가는 rmse, mae로 수행.\n",
    "gs.fit(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.3301601942969134\n",
      "{'n_epochs': 60, 'n_factors': 50}\n"
     ]
    }
   ],
   "source": [
    "#최고 RMSE Evaluation 점수와 그때의 하이퍼 파라미터는?\n",
    "print(gs.best_score['rmse'])\n",
    "print(gs.best_params['rmse'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}