notebooks/2-Embeddings Creation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "a2f0c849",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import Necessary Libraries\n",
    "import pandas as pd\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.ensemble import GradientBoostingClassifier, VotingClassifier\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.model_selection import cross_val_score\n",
    "import joblib\n",
    "import numpy as np\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "7d0d6e6f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load the preprocessed data\n",
    "preprocessed_data = pd.read_csv('preprocessed_data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "78099760",
   "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>Transaction</th>\n",
       "      <th>Tag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>User10</td>\n",
       "      <td>mg bangalore</td>\n",
       "      <td>Medical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>User3</td>\n",
       "      <td>amazon bangalore</td>\n",
       "      <td>Shopping</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>User6</td>\n",
       "      <td>ravi medicine bangalore</td>\n",
       "      <td>Medical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>User6</td>\n",
       "      <td>columbia asia bangalore</td>\n",
       "      <td>Medical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>User9</td>\n",
       "      <td>meghna foods bangalore</td>\n",
       "      <td>Food</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     User              Transaction       Tag\n",
       "0  User10             mg bangalore   Medical\n",
       "1   User3         amazon bangalore  Shopping\n",
       "2   User6  ravi medicine bangalore   Medical\n",
       "3   User6  columbia asia bangalore   Medical\n",
       "4   User9   meghna foods bangalore      Food"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Top 5 entries\n",
    "preprocessed_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03934808",
   "metadata": {},
   "source": [
    "We maintain the list of users because we use user as a feature for the training dataset, when a new user comes in we will specify a value that is not in the user_available list,probably that will help model understand that this is a new user and he\n",
    "has no prior taggings available."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "1e989360",
   "metadata": {},
   "outputs": [],
   "source": [
    "# We maintain the list of users because we use user as a feature for the training dataset\n",
    "users_available = preprocessed_data['User'].tolist()\n",
    "# Dump these Users\n",
    "with open('users_available.pkl', 'wb') as f:\n",
    "    # Write the list to the file\n",
    "    pickle.dump(users_available, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b212308",
   "metadata": {},
   "source": [
    "For our use case we would need best performing Multilingual Embeddings. So lets head over to the MTEB leaderboard and pick one from there"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "15ee07b8",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dcf57910",
   "metadata": {},
   "source": [
    "We Have two options whether to use a large parameter model or a small parater model. (Tradeoff between Accuracy and Latency)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "2f2b7b1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Let us try with all-MiniLM-L6-v2\n",
    "from sentence_transformers import SentenceTransformer\n",
    "sentences = preprocessed_data['Transaction'].to_list()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "7685bec3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-0.04322374 -0.01268452  0.00395562 ... -0.05064136 -0.11488947\n",
      "  -0.0854646 ]\n",
      " [ 0.02242737 -0.0295487  -0.07189132 ... -0.07091721 -0.06124997\n",
      "  -0.0112023 ]\n",
      " [-0.06130331  0.01692494 -0.0566077  ... -0.06592673 -0.02916417\n",
      "  -0.02086799]\n",
      " ...\n",
      " [-0.01009273 -0.04832548 -0.04341302 ... -0.05622667 -0.06757103\n",
      "  -0.04366802]\n",
      " [ 0.04578929  0.02650515 -0.03105507 ... -0.02454238 -0.03998293\n",
      "  -0.03701015]\n",
      " [-0.01315921 -0.02102073  0.00700834 ... -0.08375841  0.00316238\n",
      "  -0.04991223]]\n"
     ]
    }
   ],
   "source": [
    "# Lets get the embeddings from the sentances\n",
    "model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')\n",
    "embeddings = model.encode(sentences)\n",
    "print(embeddings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "4d3ad88c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Concat these embeddings to the file\n",
    "embeddings_dataset = pd.concat([preprocessed_data,pd.DataFrame(embeddings)],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "129e7442",
   "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>Transaction</th>\n",
       "      <th>Tag</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>...</th>\n",
       "      <th>374</th>\n",
       "      <th>375</th>\n",
       "      <th>376</th>\n",
       "      <th>377</th>\n",
       "      <th>378</th>\n",
       "      <th>379</th>\n",
       "      <th>380</th>\n",
       "      <th>381</th>\n",
       "      <th>382</th>\n",
       "      <th>383</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>User10</td>\n",
       "      <td>mg bangalore</td>\n",
       "      <td>Medical</td>\n",
       "      <td>-0.043224</td>\n",
       "      <td>-0.012685</td>\n",
       "      <td>0.003956</td>\n",
       "      <td>0.040189</td>\n",
       "      <td>-0.085170</td>\n",
       "      <td>-0.005057</td>\n",
       "      <td>0.052504</td>\n",
       "      <td>...</td>\n",
       "      <td>0.013078</td>\n",
       "      <td>0.051690</td>\n",
       "      <td>-0.042813</td>\n",
       "      <td>-0.044552</td>\n",
       "      <td>-0.048992</td>\n",
       "      <td>-0.001867</td>\n",
       "      <td>0.028403</td>\n",
       "      <td>-0.050641</td>\n",
       "      <td>-0.114889</td>\n",
       "      <td>-0.085465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>User3</td>\n",
       "      <td>amazon bangalore</td>\n",
       "      <td>Shopping</td>\n",
       "      <td>0.022427</td>\n",
       "      <td>-0.029549</td>\n",
       "      <td>-0.071891</td>\n",
       "      <td>0.036993</td>\n",
       "      <td>-0.008657</td>\n",
       "      <td>0.041752</td>\n",
       "      <td>-0.044857</td>\n",
       "      <td>...</td>\n",
       "      <td>0.010486</td>\n",
       "      <td>0.026003</td>\n",
       "      <td>-0.066787</td>\n",
       "      <td>-0.051076</td>\n",
       "      <td>-0.027117</td>\n",
       "      <td>0.111215</td>\n",
       "      <td>0.089663</td>\n",
       "      <td>-0.070917</td>\n",
       "      <td>-0.061250</td>\n",
       "      <td>-0.011202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>User6</td>\n",
       "      <td>ravi medicine bangalore</td>\n",
       "      <td>Medical</td>\n",
       "      <td>-0.061303</td>\n",
       "      <td>0.016925</td>\n",
       "      <td>-0.056608</td>\n",
       "      <td>0.024087</td>\n",
       "      <td>-0.058576</td>\n",
       "      <td>0.024351</td>\n",
       "      <td>0.017888</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.074376</td>\n",
       "      <td>0.034535</td>\n",
       "      <td>-0.060439</td>\n",
       "      <td>-0.056292</td>\n",
       "      <td>-0.052263</td>\n",
       "      <td>0.025619</td>\n",
       "      <td>0.043099</td>\n",
       "      <td>-0.065927</td>\n",
       "      <td>-0.029164</td>\n",
       "      <td>-0.020868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>User6</td>\n",
       "      <td>columbia asia bangalore</td>\n",
       "      <td>Medical</td>\n",
       "      <td>0.033344</td>\n",
       "      <td>0.009651</td>\n",
       "      <td>-0.023548</td>\n",
       "      <td>0.033597</td>\n",
       "      <td>-0.070821</td>\n",
       "      <td>0.023675</td>\n",
       "      <td>-0.006321</td>\n",
       "      <td>...</td>\n",
       "      <td>0.032060</td>\n",
       "      <td>0.048686</td>\n",
       "      <td>-0.072227</td>\n",
       "      <td>-0.026805</td>\n",
       "      <td>-0.019570</td>\n",
       "      <td>0.109888</td>\n",
       "      <td>0.026178</td>\n",
       "      <td>-0.061520</td>\n",
       "      <td>-0.093940</td>\n",
       "      <td>-0.021665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>User9</td>\n",
       "      <td>meghna foods bangalore</td>\n",
       "      <td>Food</td>\n",
       "      <td>-0.036506</td>\n",
       "      <td>-0.024542</td>\n",
       "      <td>-0.048879</td>\n",
       "      <td>0.048593</td>\n",
       "      <td>-0.079268</td>\n",
       "      <td>0.097876</td>\n",
       "      <td>-0.055953</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.027787</td>\n",
       "      <td>0.078949</td>\n",
       "      <td>-0.075628</td>\n",
       "      <td>0.023413</td>\n",
       "      <td>0.014441</td>\n",
       "      <td>0.013229</td>\n",
       "      <td>0.010613</td>\n",
       "      <td>-0.029405</td>\n",
       "      <td>-0.041154</td>\n",
       "      <td>-0.031360</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 387 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     User              Transaction       Tag         0         1         2  \\\n",
       "0  User10             mg bangalore   Medical -0.043224 -0.012685  0.003956   \n",
       "1   User3         amazon bangalore  Shopping  0.022427 -0.029549 -0.071891   \n",
       "2   User6  ravi medicine bangalore   Medical -0.061303  0.016925 -0.056608   \n",
       "3   User6  columbia asia bangalore   Medical  0.033344  0.009651 -0.023548   \n",
       "4   User9   meghna foods bangalore      Food -0.036506 -0.024542 -0.048879   \n",
       "\n",
       "          3         4         5         6  ...       374       375       376  \\\n",
       "0  0.040189 -0.085170 -0.005057  0.052504  ...  0.013078  0.051690 -0.042813   \n",
       "1  0.036993 -0.008657  0.041752 -0.044857  ...  0.010486  0.026003 -0.066787   \n",
       "2  0.024087 -0.058576  0.024351  0.017888  ... -0.074376  0.034535 -0.060439   \n",
       "3  0.033597 -0.070821  0.023675 -0.006321  ...  0.032060  0.048686 -0.072227   \n",
       "4  0.048593 -0.079268  0.097876 -0.055953  ... -0.027787  0.078949 -0.075628   \n",
       "\n",
       "        377       378       379       380       381       382       383  \n",
       "0 -0.044552 -0.048992 -0.001867  0.028403 -0.050641 -0.114889 -0.085465  \n",
       "1 -0.051076 -0.027117  0.111215  0.089663 -0.070917 -0.061250 -0.011202  \n",
       "2 -0.056292 -0.052263  0.025619  0.043099 -0.065927 -0.029164 -0.020868  \n",
       "3 -0.026805 -0.019570  0.109888  0.026178 -0.061520 -0.093940 -0.021665  \n",
       "4  0.023413  0.014441  0.013229  0.010613 -0.029405 -0.041154 -0.031360  \n",
       "\n",
       "[5 rows x 387 columns]"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Top 5 entries\n",
    "embeddings_dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "8c7b47e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Replace User with '' to make it int\n",
    "embeddings_dataset['User_id'] = embeddings_dataset['User'].str.replace('User','')\n",
    "# Change the datatype of the UserId from string to the integer\n",
    "embeddings_dataset['User_id'] = embeddings_dataset['User_id'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "8346024d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Column names can't be of mixed datatype therefore change it to string \n",
    "num_dimensions = 384\n",
    "column_mapping = {i: f'Feature{i}' for i in range(num_dimensions+1)}\n",
    "embeddings_dataset = embeddings_dataset.rename(columns=column_mapping)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "6482c53e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Prepare your featureset and the target variable\n",
    "X = embeddings_dataset.drop(columns=['Tag','User','Transaction'])\n",
    "y = embeddings_dataset['Tag']\n",
    "# Split your data into training and test sets using stratified split on the user_id. We want all users to be in the training set.\n",
    "# Will discuss why we want such a behaviour. \n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,stratify =embeddings_dataset['User_id'] ,random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "b15d819f",
   "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>Feature0</th>\n",
       "      <th>Feature1</th>\n",
       "      <th>Feature2</th>\n",
       "      <th>Feature3</th>\n",
       "      <th>Feature4</th>\n",
       "      <th>Feature5</th>\n",
       "      <th>Feature6</th>\n",
       "      <th>Feature7</th>\n",
       "      <th>Feature8</th>\n",
       "      <th>Feature9</th>\n",
       "      <th>...</th>\n",
       "      <th>Feature375</th>\n",
       "      <th>Feature376</th>\n",
       "      <th>Feature377</th>\n",
       "      <th>Feature378</th>\n",
       "      <th>Feature379</th>\n",
       "      <th>Feature380</th>\n",
       "      <th>Feature381</th>\n",
       "      <th>Feature382</th>\n",
       "      <th>Feature383</th>\n",
       "      <th>User_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>466</th>\n",
       "      <td>-0.061607</td>\n",
       "      <td>0.021109</td>\n",
       "      <td>-0.029507</td>\n",
       "      <td>0.050608</td>\n",
       "      <td>0.001134</td>\n",
       "      <td>0.005973</td>\n",
       "      <td>0.040485</td>\n",
       "      <td>0.028323</td>\n",
       "      <td>-0.034021</td>\n",
       "      <td>0.046917</td>\n",
       "      <td>...</td>\n",
       "      <td>0.055946</td>\n",
       "      <td>-0.067546</td>\n",
       "      <td>-0.101885</td>\n",
       "      <td>-0.006615</td>\n",
       "      <td>0.073394</td>\n",
       "      <td>0.066049</td>\n",
       "      <td>-0.001069</td>\n",
       "      <td>-0.073724</td>\n",
       "      <td>-0.023892</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129</th>\n",
       "      <td>-0.025332</td>\n",
       "      <td>-0.051232</td>\n",
       "      <td>-0.100755</td>\n",
       "      <td>0.055770</td>\n",
       "      <td>-0.009053</td>\n",
       "      <td>-0.016233</td>\n",
       "      <td>0.007624</td>\n",
       "      <td>0.025452</td>\n",
       "      <td>-0.021775</td>\n",
       "      <td>-0.015170</td>\n",
       "      <td>...</td>\n",
       "      <td>0.064789</td>\n",
       "      <td>-0.051988</td>\n",
       "      <td>-0.059050</td>\n",
       "      <td>-0.063797</td>\n",
       "      <td>0.078680</td>\n",
       "      <td>0.077889</td>\n",
       "      <td>-0.029503</td>\n",
       "      <td>-0.051294</td>\n",
       "      <td>-0.037061</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>940</th>\n",
       "      <td>-0.013159</td>\n",
       "      <td>-0.021021</td>\n",
       "      <td>0.007008</td>\n",
       "      <td>-0.022461</td>\n",
       "      <td>0.002843</td>\n",
       "      <td>0.021604</td>\n",
       "      <td>0.028168</td>\n",
       "      <td>0.084881</td>\n",
       "      <td>-0.046810</td>\n",
       "      <td>-0.008587</td>\n",
       "      <td>...</td>\n",
       "      <td>0.019474</td>\n",
       "      <td>-0.091984</td>\n",
       "      <td>-0.030931</td>\n",
       "      <td>-0.078524</td>\n",
       "      <td>0.062954</td>\n",
       "      <td>-0.054538</td>\n",
       "      <td>-0.083758</td>\n",
       "      <td>0.003162</td>\n",
       "      <td>-0.049912</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>730</th>\n",
       "      <td>-0.014338</td>\n",
       "      <td>0.036519</td>\n",
       "      <td>-0.030587</td>\n",
       "      <td>-0.043091</td>\n",
       "      <td>-0.097023</td>\n",
       "      <td>0.038005</td>\n",
       "      <td>-0.047036</td>\n",
       "      <td>-0.017092</td>\n",
       "      <td>-0.057033</td>\n",
       "      <td>0.007419</td>\n",
       "      <td>...</td>\n",
       "      <td>0.151159</td>\n",
       "      <td>-0.095012</td>\n",
       "      <td>-0.085966</td>\n",
       "      <td>-0.076053</td>\n",
       "      <td>0.035824</td>\n",
       "      <td>0.032255</td>\n",
       "      <td>-0.076523</td>\n",
       "      <td>-0.043340</td>\n",
       "      <td>-0.072152</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210</th>\n",
       "      <td>0.020997</td>\n",
       "      <td>-0.052240</td>\n",
       "      <td>-0.077261</td>\n",
       "      <td>-0.066575</td>\n",
       "      <td>0.055326</td>\n",
       "      <td>0.087589</td>\n",
       "      <td>-0.029689</td>\n",
       "      <td>0.046233</td>\n",
       "      <td>-0.017779</td>\n",
       "      <td>-0.011467</td>\n",
       "      <td>...</td>\n",
       "      <td>0.066864</td>\n",
       "      <td>-0.018092</td>\n",
       "      <td>-0.051249</td>\n",
       "      <td>-0.065001</td>\n",
       "      <td>0.036361</td>\n",
       "      <td>0.145354</td>\n",
       "      <td>0.001424</td>\n",
       "      <td>-0.049648</td>\n",
       "      <td>-0.069986</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 385 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Feature0  Feature1  Feature2  Feature3  Feature4  Feature5  Feature6  \\\n",
       "466 -0.061607  0.021109 -0.029507  0.050608  0.001134  0.005973  0.040485   \n",
       "129 -0.025332 -0.051232 -0.100755  0.055770 -0.009053 -0.016233  0.007624   \n",
       "940 -0.013159 -0.021021  0.007008 -0.022461  0.002843  0.021604  0.028168   \n",
       "730 -0.014338  0.036519 -0.030587 -0.043091 -0.097023  0.038005 -0.047036   \n",
       "210  0.020997 -0.052240 -0.077261 -0.066575  0.055326  0.087589 -0.029689   \n",
       "\n",
       "     Feature7  Feature8  Feature9  ...  Feature375  Feature376  Feature377  \\\n",
       "466  0.028323 -0.034021  0.046917  ...    0.055946   -0.067546   -0.101885   \n",
       "129  0.025452 -0.021775 -0.015170  ...    0.064789   -0.051988   -0.059050   \n",
       "940  0.084881 -0.046810 -0.008587  ...    0.019474   -0.091984   -0.030931   \n",
       "730 -0.017092 -0.057033  0.007419  ...    0.151159   -0.095012   -0.085966   \n",
       "210  0.046233 -0.017779 -0.011467  ...    0.066864   -0.018092   -0.051249   \n",
       "\n",
       "     Feature378  Feature379  Feature380  Feature381  Feature382  Feature383  \\\n",
       "466   -0.006615    0.073394    0.066049   -0.001069   -0.073724   -0.023892   \n",
       "129   -0.063797    0.078680    0.077889   -0.029503   -0.051294   -0.037061   \n",
       "940   -0.078524    0.062954   -0.054538   -0.083758    0.003162   -0.049912   \n",
       "730   -0.076053    0.035824    0.032255   -0.076523   -0.043340   -0.072152   \n",
       "210   -0.065001    0.036361    0.145354    0.001424   -0.049648   -0.069986   \n",
       "\n",
       "     User_id  \n",
       "466       13  \n",
       "129       14  \n",
       "940       12  \n",
       "730       14  \n",
       "210        1  \n",
       "\n",
       "[5 rows x 385 columns]"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}