{
"cells": [
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"\n",
"# read the train_data.csv, test_data.csv and videos_metadata.csv\n",
"train_data = pd.read_csv(\"train_data.csv\")\n",
"test_data = pd.read_csv(\"test_data.csv\")\n",
"videos_metadata = pd.read_csv(\"videos_metadata.csv\")\n",
"\n",
"# the file users_data.csv contains values that pandas function read_csv could not decode, so we will read the file \n",
"# users_data.csv manually using the python standard function open()\n",
"\n",
"users_metadata_ = []\n",
"with open(\"users_metadata.csv\") as file:\n",
" content = file.readlines()\n",
" for row in content:\n",
"    row = row.strip(\"\\n\").split(\",\")\n",
"    users_metadata_.append(row)\n",
"    \n",
"users_metadata = pd.DataFrame(users_metadata_[1:], columns=users_metadata_[0])\n",
"users_metadata[\"uploads\"] = pd.to_numeric(users_metadata[\"uploads\"])\n",
"users_metadata[\"watches\"] = pd.to_numeric(users_metadata[\"watches\"])\n",
"users_metadata[\"friends\"] = pd.to_numeric(users_metadata[\"friends\"])\n",
"\n",
"# now we have all datasets prepared"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [],
"source": [
"def encode_string(dataframe, column_name):\n",
" \"\"\"A function that accepts a dataframe and the name of the string attribute. \n",
" It returns a dataframe with one-hot encoded attribute \"\"\"\n",
" \n",
" str_column = dataframe[column_name].values\n",
" unique_values = list(set(str_column))\n",
" onehot_attributes = []\n",
"\n",
" for val in str_column: # iterate over string column\n",
"    one_hot = [0 for l in unique_values]\n",
"    for ind, val_ in enumerate(unique_values):\n",
"          if val_ == val: \n",
"             one_hot[ind] = 1\n",
"    onehot_attributes.append(one_hot)\n",
" \n",
" \n",
" # make a new dataframe with onehot encoded attributes \n",
" df_ = pd.DataFrame(onehot_attributes)\n",
" df_.columns = unique_values\n",
" new_df = pd.merge(dataframe, df_, left_index=True, right_index=True)\n",
" # drop the column with the string attribute\n",
" new_df.drop(column_name, axis = 1, inplace = True)\n",
" return new_df"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# First look into the data"
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
   "train data length: 26901\n",
   "test data length: 99500\n",
   "videos metadata length: 197317\n",
   "users metadata length: 1062324\n"
]
}
],
"source": [
"print(\"train data length:\", len(train_data))\n",
"print(\"test data length:\", len(test_data))\n",
"print(\"videos metadata length:\", len(videos_metadata))\n",
"print(\"users metadata length:\", len(users_metadata))"
]
},
{
"cell_type": "code",
"execution_count": 30,
"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>video_id</th>\n",
   "    <th>uploader</th>\n",
   "    <th>age</th>\n",
   "    <th>category</th>\n",
   "    <th>length</th>\n",
   "    <th>views</th>\n",
   "    <th>rate</th>\n",
   "    <th>ratings</th>\n",
   "    <th>comments</th>\n",
   "    <th>1</th>\n",
   "    <th>...</th>\n",
   "    <th>11</th>\n",
   "    <th>12</th>\n",
   "    <th>13</th>\n",
   "    <th>14</th>\n",
   "    <th>15</th>\n",
   "    <th>16</th>\n",
   "    <th>17</th>\n",
   "    <th>18</th>\n",
   "    <th>19</th>\n",
   "    <th>20</th>\n",
   " </tr>\n",
   " </thead>\n",
   " <tbody>\n",
   " <tr>\n",
   "    <th>0</th>\n",
   "    <td>LKh7zAJ4nwo</td>\n",
   "    <td>TheReceptionist</td>\n",
   "    <td>653.0</td>\n",
   "    <td>Entertainment</td>\n",
   "    <td>424.0</td>\n",
   "    <td>13021.0</td>\n",
   "    <td>4.34</td>\n",
   "    <td>1305.0</td>\n",
   "    <td>744.0</td>\n",
   "    <td>NaN</td>\n",
   "    <td>...</td>\n",
   "    <td>DjdA-5oKYFQ</td>\n",
   "    <td>vmdPOOd6cxI</td>\n",
   "    <td>uX81lMev6_o</td>\n",
   "    <td>NaN</td>\n",
   "    <td>W1Uo5DQTtzc</td>\n",
   "    <td>heqocRij5P0</td>\n",
   "    <td>2ZgXx72XmoE</td>\n",
   "    <td>1tUDzOp10pk</td>\n",
   "    <td>DH56yrIO5nI</td>\n",
   "    <td>E-3zXq_r4w0</td>\n",
   " </tr>\n",
   " <tr>\n",
   "    <th>1</th>\n",
   "    <td>7D0Mf4Kn4Xk</td>\n",
   "    <td>periurban</td>\n",
   "    <td>583.0</td>\n",
   "    <td>Music</td>\n",
   "    <td>201.0</td>\n",
   "    <td>6508.0</td>\n

Consider the following recursive algorithm: Unknown(A,i,j); {...