06-6. Adam, tanh+relu.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import LabelEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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_id</th>\n",
       "      <th>username</th>\n",
       "      <th>date</th>\n",
       "      <th>text</th>\n",
       "      <th>look</th>\n",
       "      <th>smell</th>\n",
       "      <th>taste</th>\n",
       "      <th>feel</th>\n",
       "      <th>overall</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>271781</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>2017-03-17</td>\n",
       "      <td>750 ml bottle, 2016 vintage, bottle #304 of...</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>125646</td>\n",
       "      <td>_dirty_</td>\n",
       "      <td>2017-12-21</td>\n",
       "      <td></td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>125646</td>\n",
       "      <td>CJDUBYA</td>\n",
       "      <td>2017-12-21</td>\n",
       "      <td></td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>125646</td>\n",
       "      <td>GratefulBeerGuy</td>\n",
       "      <td>2017-12-20</td>\n",
       "      <td>0% 16 oz can. Funny story: As I finally wal...</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.75</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>125646</td>\n",
       "      <td>LukeGude</td>\n",
       "      <td>2017-12-20</td>\n",
       "      <td>Classic TH NEIPA. Overflowing head and bouq...</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.50</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.25</td>\n",
       "      <td>4.31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   beer_id         username        date  \\\n",
       "0   271781     bluejacket74  2017-03-17   \n",
       "1   125646          _dirty_  2017-12-21   \n",
       "2   125646          CJDUBYA  2017-12-21   \n",
       "3   125646  GratefulBeerGuy  2017-12-20   \n",
       "4   125646         LukeGude  2017-12-20   \n",
       "\n",
       "                                                text  look  smell  taste  \\\n",
       "0     750 ml bottle, 2016 vintage, bottle #304 of...  4.00   4.00   4.00   \n",
       "1                                                     4.50   4.50   4.50   \n",
       "2                                                     4.75   4.75   4.75   \n",
       "3     0% 16 oz can. Funny story: As I finally wal...  4.75   4.75   4.50   \n",
       "4     Classic TH NEIPA. Overflowing head and bouq...  4.25   4.50   4.25   \n",
       "\n",
       "   feel  overall  score  \n",
       "0  4.25     4.00   4.03  \n",
       "1  4.50     4.50   4.50  \n",
       "2  4.75     4.75   4.75  \n",
       "3  4.50     4.50   4.58  \n",
       "4  4.25     4.25   4.31  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('D:/stat_team/project/beers-breweries-and-beer-reviews/reviews.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>beerid</th>\n",
       "      <th>userid</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>271781</td>\n",
       "      <td>bluejacket74</td>\n",
       "      <td>4.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>125646</td>\n",
       "      <td>_dirty_</td>\n",
       "      <td>4.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>125646</td>\n",
       "      <td>CJDUBYA</td>\n",
       "      <td>4.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>125646</td>\n",
       "      <td>GratefulBeerGuy</td>\n",
       "      <td>4.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>125646</td>\n",
       "      <td>LukeGude</td>\n",
       "      <td>4.31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   beerid           userid  rating\n",
       "0  271781     bluejacket74    4.03\n",
       "1  125646          _dirty_    4.50\n",
       "2  125646          CJDUBYA    4.75\n",
       "3  125646  GratefulBeerGuy    4.58\n",
       "4  125646         LukeGude    4.31"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#잠깐 전처리\n",
    "\n",
    "cols = ['beer_id','username','score']\n",
    "df = df[cols]\n",
    "df.rename(columns = {'beer_id':'beerid','username':'userid','score':'rating'}, inplace = True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(164935, 309542, 1.0, 5.0)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#딥러닝에 이용할 컬럼 변경: user, beer, rating\n",
    "\n",
    "user_enc = LabelEncoder()\n",
    "df['user'] = user_enc.fit_transform(df['userid'].tolist())\n",
    "n_users = df['user'].nunique()\n",
    "\n",
    "item_enc = LabelEncoder()\n",
    "df['beer'] = item_enc.fit_transform(df['beerid'].values)\n",
    "n_beers = df['beer'].nunique()\n",
    "\n",
    "df['rating'] = df['rating'].values.astype(np.float32)\n",
    "min_rating = min(df['rating'])\n",
    "max_rating = max(df['rating'])\n",
    "\n",
    "n_users, n_beers, min_rating, max_rating #n_users, n_beers: 유니크한 값 개수, rating의 최솟값과 최댓값."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>user</th>\n",
       "      <th>beer</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>102213</td>\n",
       "      <td>231507</td>\n",
       "      <td>4.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>94429</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15635</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>34826</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>55007</td>\n",
       "      <td>112229</td>\n",
       "      <td>4.31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     user    beer  rating\n",
       "0  102213  231507    4.03\n",
       "1   94429  112229    4.50\n",
       "2   15635  112229    4.75\n",
       "3   34826  112229    4.58\n",
       "4   55007  112229    4.31"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2=df[['user','beer','rating']] \n",
    "df2.head() #완료된 데이터셋. 여기서 'user','beer','rating'만 사용"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((7258502, 2), (1814626, 2), (7258502,), (1814626,))"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#train, test 셋 분할: test size는 0.2, random_state는 임의로 지정\n",
    "X = df[['user', 'beer']].values\n",
    "y = df['rating'].values\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "\n",
    "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#딥러닝 모델을 만들 때 필요한 또다른 변수는: 잠재요인의 개수를 정하는 것!\n",
    "#잠재요인의 개수는 원하는 대로 설정하면 되지만, 주의할 것은 user와 beer 모두에 대해 같은 사이즈를 줘야 한다는 점.\n",
    "\n",
    "#우리는 50개로 간다!(X에 대해서만 적용)\n",
    "n_factors = 50\n",
    "\n",
    "#X를 array로 만들기\n",
    "X_train_array = [X_train[:, 0], X_train[:, 1]]\n",
    "X_test_array = [X_test[:, 0], X_test[:, 1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "beers = pd.read_csv('D:/stat_team/project/beers-breweries-and-beer-reviews/beers.csv')\n",
    "\n",
    "## create a dictionary out of beerid, beer original title\n",
    "beers_dict = beers.set_index('id')['name'].to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "#keras 라이브러리 로드\n",
    "from keras.models import Model\n",
    "from keras.layers import Input, Reshape, Dot\n",
    "from keras.layers import Embedding, Flatten, Dense, Dropout, concatenate, multiply, Input\n",
    "from keras.optimizers import Adam\n",
    "from keras.regularizers import l2\n",
    "from keras.layers import Add, Activation, Lambda, Concatenate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "##시각화(각 epoch마다의 MSE, MAE 값 비교)\n",
    "\n",
    "def plot_history(hist):\n",
    "    \n",
    "    plt.figure(figsize=(10,15))\n",
    "\n",
    "    plt.subplot(2,1,1)\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Mean Abs Error [rating]')\n",
    "    plt.plot(hist['epoch'], hist['mae'], label='Train Error')\n",
    "    plt.plot(hist['epoch'], hist['val_mae'],label = 'Test Error')\n",
    "    plt.legend()\n",
    "\n",
    "    plt.subplot(2,1,2)\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Mean Square Error [$rating^2$]')\n",
    "    plt.plot(hist['epoch'], hist['mse'], label='Train Error')\n",
    "    plt.plot(hist['epoch'], hist['val_mse'],label = 'Test Error')\n",
    "    plt.legend()\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 옵티마이저 Adam / tanh->relu / batch_size = 1000 / epoch = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "#원래 함수(수정본2)\n",
    "##bias가 있음, 아주 단순한 버전. 그냥 R_hat = u*m으로 연산한다. 최저 rating과 최고 rating을 이용한다.\n",
    "##layer 2개를 추가하였고, activation 함수 2개는: relu와 sigmoid\n",
    "\n",
    "class EmbeddingLayer:\n",
    "    def __init__(self, n_items, n_factors):\n",
    "        self.n_items = n_items\n",
    "        self.n_factors = n_factors\n",
    "    \n",
    "    def __call__(self, x):\n",
    "        x = Embedding(self.n_items, self.n_factors, embeddings_initializer='he_normal')(x) \n",
    "        x = Reshape((self.n_factors,))(x)\n",
    "        return x\n",
    "\n",
    "def RecommenderV5_1(n_users, n_beers, n_factors):\n",
    "    \n",
    "    #input: user, beer\n",
    "    user = Input(shape=(1,))\n",
    "    u = EmbeddingLayer(n_users, n_factors)(user) #u는 user * n_factor\n",
    "    ub = EmbeddingLayer(n_users, 1)(user) #user_bias\n",
    "    \n",
    "    beer = Input(shape=(1,))\n",
    "    b = EmbeddingLayer(n_beers, n_factors)(beer) #b은 beer * n_factor\n",
    "    bb = EmbeddingLayer(n_beers, 1)(beer) #beer_bias\n",
    "    \n",
    "    #input 연산- activation 2가지를 추가. tanh과 relu\n",
    "    x = Concatenate()([u, b])\n",
    "    \n",
    "    x = Dense(10, kernel_initializer='he_normal')(x)\n",
    "    x = Activation('tanh')(x)\n",
    "    \n",
    "    x = Dense(1, kernel_initializer='he_normal')(x)\n",
    "    x = Activation('relu')(x)\n",
    "    x = Lambda(lambda x: x * (max_rating - min_rating) + min_rating)(x)\n",
    "\n",
    "\n",
    "    #최종 모델 만들기(input은 user, beer 2개를 받고, output은 x를 반환.)\n",
    "    model = Model(inputs=[user, beer], outputs=x)\n",
    "    opt = Adam(lr=0.001) #learning rate(학습률) 설정. #optimizer인 최적화 방법은 Adam을 씀.\n",
    "    model.compile(loss=['mse'], metrics = ['mae','mse'] ,optimizer=opt) #loss와 함께 출력할 mae, mse는 metrics로 설정.\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_4\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_7 (InputLayer)            (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_8 (InputLayer)            (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "embedding_13 (Embedding)        (None, 1, 50)        8246750     input_7[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "embedding_15 (Embedding)        (None, 1, 50)        15477100    input_8[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "reshape_13 (Reshape)            (None, 50)           0           embedding_13[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "reshape_15 (Reshape)            (None, 50)           0           embedding_15[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_4 (Concatenate)     (None, 100)          0           reshape_13[0][0]                 \n",
      "                                                                 reshape_15[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_8 (Dense)                 (None, 10)           1010        concatenate_4[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_8 (Activation)       (None, 10)           0           dense_8[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_9 (Dense)                 (None, 1)            11          activation_8[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_9 (Activation)       (None, 1)            0           dense_9[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "lambda_4 (Lambda)               (None, 1)            0           activation_9[0][0]               \n",
      "==================================================================================================\n",
      "Total params: 23,724,871\n",
      "Trainable params: 23,724,871\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "re5model_1 = RecommenderV5_1(n_users, n_beers, n_factors)\n",
    "re5model_1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7258502 samples, validate on 1814626 samples\n",
      "Epoch 1/10\n",
      "7258502/7258502 [==============================] - 4433s 611us/step - loss: 0.2316 - mae: 0.3324 - mse: 0.2316 - val_loss: 0.1838 - val_mae: 0.3036 - val_mse: 0.1838\n",
      "Epoch 2/10\n",
      "7258502/7258502 [==============================] - 4781s 659us/step - loss: 0.1737 - mae: 0.2942 - mse: 0.1737 - val_loss: 0.1785 - val_mae: 0.2989 - val_mse: 0.1785\n",
      "Epoch 3/10\n",
      "7258502/7258502 [==============================] - 4354s 600us/step - loss: 0.1646 - mae: 0.2855 - mse: 0.1646 - val_loss: 0.1766 - val_mae: 0.2941 - val_mse: 0.1766\n",
      "Epoch 4/10\n",
      "7258502/7258502 [==============================] - 4062s 560us/step - loss: 0.1578 - mae: 0.2789 - mse: 0.1578 - val_loss: 0.1767 - val_mae: 0.2959 - val_mse: 0.1767\n",
      "Epoch 5/10\n",
      "7258502/7258502 [==============================] - 4060s 559us/step - loss: 0.1516 - mae: 0.2731 - mse: 0.1516 - val_loss: 0.1774 - val_mae: 0.2966 - val_mse: 0.1774\n",
      "Epoch 6/10\n",
      "7258502/7258502 [==============================] - 4073s 561us/step - loss: 0.1458 - mae: 0.2675 - mse: 0.1458 - val_loss: 0.1785 - val_mae: 0.2966 - val_mse: 0.1785\n",
      "Epoch 7/10\n",
      "7258502/7258502 [==============================] - 4054s 558us/step - loss: 0.1408 - mae: 0.2628 - mse: 0.1408 - val_loss: 0.1799 - val_mae: 0.2993 - val_mse: 0.1799\n",
      "Epoch 8/10\n",
      "7258502/7258502 [==============================] - 4066s 560us/step - loss: 0.1369 - mae: 0.2588 - mse: 0.1369 - val_loss: 0.1814 - val_mae: 0.2991 - val_mse: 0.1814\n",
      "Epoch 9/10\n",
      "7258502/7258502 [==============================] - 4045s 557us/step - loss: 0.1335 - mae: 0.2555 - mse: 0.1335 - val_loss: 0.1826 - val_mae: 0.3008 - val_mse: 0.1826\n",
      "Epoch 10/10\n",
      "7236000/7258502 [============================>.] - ETA: 12s - loss: 0.1308 - mae: 0.2527 - mse: 0.1308"
     ]
    }
   ],
   "source": [
    "## fit model\n",
    "import numpy as np\n",
    "seed = 7\n",
    "np.random.seed(seed)\n",
    "history_mf5_1 = re5model_1.fit(x=X_train_array, y=y_train, batch_size=1000, epochs=10,\n",
    "                    verbose=1, validation_data=(X_test_array, y_test)) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "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>val_loss</th>\n",
       "      <th>val_mae</th>\n",
       "      <th>val_mse</th>\n",
       "      <th>loss</th>\n",
       "      <th>mae</th>\n",
       "      <th>mse</th>\n",
       "      <th>epoch</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.183785</td>\n",
       "      <td>0.303635</td>\n",
       "      <td>0.183784</td>\n",
       "      <td>0.231564</td>\n",
       "      <td>0.332360</td>\n",
       "      <td>0.231563</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.178545</td>\n",
       "      <td>0.298941</td>\n",
       "      <td>0.178545</td>\n",
       "      <td>0.173659</td>\n",
       "      <td>0.294210</td>\n",
       "      <td>0.173659</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.176583</td>\n",
       "      <td>0.294141</td>\n",
       "      <td>0.176583</td>\n",
       "      <td>0.164562</td>\n",
       "      <td>0.285464</td>\n",
       "      <td>0.164562</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.176681</td>\n",
       "      <td>0.295940</td>\n",
       "      <td>0.176681</td>\n",
       "      <td>0.157807</td>\n",
       "      <td>0.278924</td>\n",
       "      <td>0.157807</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.177378</td>\n",
       "      <td>0.296558</td>\n",
       "      <td>0.177378</td>\n",
       "      <td>0.151592</td>\n",
       "      <td>0.273080</td>\n",
       "      <td>0.151592</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.178530</td>\n",
       "      <td>0.296585</td>\n",
       "      <td>0.178530</td>\n",
       "      <td>0.145783</td>\n",
       "      <td>0.267503</td>\n",
       "      <td>0.145783</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.179940</td>\n",
       "      <td>0.299311</td>\n",
       "      <td>0.179941</td>\n",
       "      <td>0.140834</td>\n",
       "      <td>0.262758</td>\n",
       "      <td>0.140834</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.181439</td>\n",
       "      <td>0.299128</td>\n",
       "      <td>0.181439</td>\n",
       "      <td>0.136851</td>\n",
       "      <td>0.258804</td>\n",
       "      <td>0.136851</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.182640</td>\n",
       "      <td>0.300811</td>\n",
       "      <td>0.182640</td>\n",
       "      <td>0.133521</td>\n",
       "      <td>0.255483</td>\n",
       "      <td>0.133521</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.183617</td>\n",
       "      <td>0.302124</td>\n",
       "      <td>0.183617</td>\n",
       "      <td>0.130816</td>\n",
       "      <td>0.252755</td>\n",
       "      <td>0.130816</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   val_loss   val_mae   val_mse      loss       mae       mse  epoch\n",
       "0  0.183785  0.303635  0.183784  0.231564  0.332360  0.231563      0\n",
       "1  0.178545  0.298941  0.178545  0.173659  0.294210  0.173659      1\n",
       "2  0.176583  0.294141  0.176583  0.164562  0.285464  0.164562      2\n",
       "3  0.176681  0.295940  0.176681  0.157807  0.278924  0.157807      3\n",
       "4  0.177378  0.296558  0.177378  0.151592  0.273080  0.151592      4\n",
       "5  0.178530  0.296585  0.178530  0.145783  0.267503  0.145783      5\n",
       "6  0.179940  0.299311  0.179941  0.140834  0.262758  0.140834      6\n",
       "7  0.181439  0.299128  0.181439  0.136851  0.258804  0.136851      7\n",
       "8  0.182640  0.300811  0.182640  0.133521  0.255483  0.133521      8\n",
       "9  0.183617  0.302124  0.183617  0.130816  0.252755  0.130816      9"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## show loss & mae & mse at each epoch\n",
    "hist5_1 = pd.DataFrame(history_mf5_1.history)\n",
    "hist5_1['epoch'] = history_mf5_1.epoch\n",
    "hist5_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1814626/1814626 [==============================] - 3s 2us/step\n",
      "## evaluation loss and_metrics ##\n",
      "[0.18361665110496375, 0.30212438106536865, 0.18361686170101166]\n"
     ]
    }
   ],
   "source": [
    "#모델 평가하기; model.evaluate() #test set에 평가했을 때, 맨 마지막 epoch의 val_loss, val_mae, val_mse\n",
    "loss_and_metrics_1 = re5model_1.evaluate(X_test_array, y_test, batch_size=1000)\n",
    "print('## evaluation loss and_metrics ##')\n",
    "print(loss_and_metrics_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(309542, 50)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## get weights of the beers embedding matrix\n",
    "beer_embedding_weights_1 = re5model_1.layers[3].get_weights()[0] #3은 beer-embedding : layer 확인!\n",
    "beer_embedding_weights_1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3, 309542)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## import PCA\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "pca = PCA(n_components = 3) ## use 3 components\n",
    "beer_embedding_weights_t_1 = np.transpose(beer_embedding_weights_1) ## pass the transpose of the embedding matrix\n",
    "beer_pca_1 = pca.fit(beer_embedding_weights_t_1) ## fit\n",
    "\n",
    "## display the resulting matrix dimensions\n",
    "beer_pca_1.components_.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1080 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_history(hist5_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 4min 44s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "unbeers = get_unbeer_keras(df2, 190)\n",
    "recomm_beer_by_keras(190, re5model_1, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(309542, 50)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## get weights of the beers embedding matrix\n",
    "beer_embedding_weights_1 = re5model_1.layers[3].get_weights()[0] #3은 beer-embedding : layer 확인!\n",
    "beer_embedding_weights_1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3, 309542)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## import PCA\n",
    "pca = PCA(n_components = 3) ## use 3 components\n",
    "beer_embedding_weights_t_1 = np.transpose(beer_embedding_weights_1) ## pass the transpose of the embedding matrix\n",
    "beer_pca_1 = pca.fit(beer_embedding_weights_t_1) ## fit\n",
    "\n",
    "## display the resulting matrix dimensions\n",
    "beer_pca_1.components_.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.46143547, 0.10642262, 0.0560739 ], dtype=float32)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## display the variance explained by the 3 components\n",
    "beer_pca_1.explained_variance_ratio_\n",
    " #variance 가 아주 낮으면 좋은 해석을 기대하기 어렵다.\n",
    "    #첫번째 PC 분산은 약 얼마? 78%로, dropout 규제를 넣었는데도 성능이 좋다!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "from operator import itemgetter\n",
    "\n",
    "## extract first PCA\n",
    "pca1 = beer_pca_1.components_[0]\n",
    "\n",
    "## get the value (pca0, beer title)\n",
    "beer_comp1 = [(f, beers_dict[i]) for f,i in zip(pca1, list(beers_dict.keys()))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0.008864303, 'Sir Phobos'),\n",
       " (0.007522378, 'Cream Ale'),\n",
       " (0.0074552475, 'Whodunnit?'),\n",
       " (0.007221045, 'The Legendary Duff Beer'),\n",
       " (0.007047164, 'Motor Oil No. 2'),\n",
       " (0.0070388354, \"Weiherer Fat Head's Hopferla IPA\"),\n",
       " (0.0070090326, 'Spendrups Premium Lager Alkoholfri'),\n",
       " (0.0069139455, 'Penn Ginger Beer'),\n",
       " (0.006913174, 'Powder'),\n",
       " (0.0068994896, 'Avalanche')]"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## beers corresponding to the highest values of pca0\n",
    "sorted(beer_comp1, key = itemgetter(0), reverse = True)[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(-0.0076245004, 'Snowman Make Like A Tree And Leaf Maple Porter'),\n",
       " (-0.0074016056, 'Scottish Ale'),\n",
       " (-0.0073216716, 'Veluwse Schavuyt Tripel'),\n",
       " (-0.007302906, 'Red Truck Porter'),\n",
       " (-0.0072260145, 'Quarters Mostly'),\n",
       " (-0.007202047, 'Pecan Porter'),\n",
       " (-0.0071978183, 'Your Energy Amber ale'),\n",
       " (-0.007190992, 'Sour Concord Grape'),\n",
       " (-0.007181778, 'One-Eyed Jack Imperial Pilsner'),\n",
       " (-0.0071377484, 'Winter Sesh')]"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## beers corresponding to the lowest values of pca0\n",
    "sorted(beer_comp1, key = itemgetter(0))[:10]"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}