618_Homework_07.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Homework 7: Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "HiF5779SXEsY",
    "outputExpanded": false
   },
   "source": [
    "## The total score for this assignment will be 100 points, consisting of:\n",
    "- 10 pt: Overall quality of spelling, grammar, puctuation, etc. of written sentences.\n",
    "- 10 pt: Codes are written in [PEP 8](https://www.python.org/dev/peps/pep-0008/) style.\n",
    "- 80 pt: Homework questions. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This homework assignment is very similar to the in-class lab exercise that you completed this week.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset is from the \"Malicious and Benign Websites\" Kaggle project at https://www.kaggle.com/xwolf12/malicious-and-benign-websites/home (you'll need to go to that URL to get a full description of the data).\n",
    "\n",
    "From the overview:\n",
    "\n",
    "The project consisted to evaluate different classification models to predict malicious and benign websites, based on application layer and network characteristics. The data were obtained by using different verified sources of benign and malicious URL's, in a low interactive client honeypot to isolate network traffic. We used additional tools to get other information, such as, server country with Whois."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Your task is to build the best classifier you can, choosing from either Random Forests or Naive Bayes classifiers (i.e. no XGBoost, SVM, Logistic Regression, etc.).\n",
    "\n",
    "You will need to split the complete dataset into training and testing datasets.  We recommend a 70%-30% split."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. Build both Random Forest (10 points) and Naive Bayes (10 points) classifiers to predict the TYPE variable.  Be sure to provide interpretations of your results.\n",
    "\n",
    "2. Experiment with changing the number of \"folds\" in your cross-validator for your Random Forest classifiers (10 points). Be sure to provide interpretations of your results.\n",
    "\n",
    "3. Tune your Random Forest model by varying hyperparamters (15 points). Be sure to provide interpretations of your results.\n",
    "\n",
    "3. Experiment with setting the appropriate values for *priors* in the Naive Bayes model (the priors should correspond to the proportions of the TYPE values) (10 points). Be sure to provide interpretations of your results.\n",
    "\n",
    "4. Report the most important feature(s) based on your Random Forest classifier (10 points). Are you surprised by the results?  Explain why.\n",
    "\n",
    "5. Summarize your findings, including which classifier and paramters gave you the best results and include an in-depth explanation of how you might improve your results.  You should also investigate what the different fields mean, so you'll need to do some research about websites and internet protocols! (15 points)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scipy as sp\n",
    "import sklearn as sk\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import cross_val_score\n",
    "import sklearn.ensemble as skens\n",
    "import sklearn.metrics as skmetric\n",
    "import sklearn.naive_bayes as sknb\n",
    "import sklearn.tree as sktree\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "sns.set(style='white', color_codes=True, font_scale=1.3)\n",
    "import sklearn.externals.six as sksix\n",
    "import IPython.display as ipd\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn import metrics\n",
    "import os\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('malicious-and-benign-websites.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "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>URL</th>\n",
       "      <th>URL_LENGTH</th>\n",
       "      <th>NUMBER_SPECIAL_CHARACTERS</th>\n",
       "      <th>CHARSET</th>\n",
       "      <th>SERVER</th>\n",
       "      <th>CONTENT_LENGTH</th>\n",
       "      <th>WHOIS_COUNTRY</th>\n",
       "      <th>WHOIS_STATEPRO</th>\n",
       "      <th>WHOIS_REGDATE</th>\n",
       "      <th>WHOIS_UPDATED_DATE</th>\n",
       "      <th>...</th>\n",
       "      <th>DIST_REMOTE_TCP_PORT</th>\n",
       "      <th>REMOTE_IPS</th>\n",
       "      <th>APP_BYTES</th>\n",
       "      <th>SOURCE_APP_PACKETS</th>\n",
       "      <th>REMOTE_APP_PACKETS</th>\n",
       "      <th>SOURCE_APP_BYTES</th>\n",
       "      <th>REMOTE_APP_BYTES</th>\n",
       "      <th>APP_PACKETS</th>\n",
       "      <th>DNS_QUERY_TIMES</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>M0_109</td>\n",
       "      <td>16</td>\n",
       "      <td>7</td>\n",
       "      <td>iso-8859-1</td>\n",
       "      <td>nginx</td>\n",
       "      <td>263.0</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>10/10/2015 18:21</td>\n",
       "      <td>None</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>700</td>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>1153</td>\n",
       "      <td>832</td>\n",
       "      <td>9</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B0_2314</td>\n",
       "      <td>16</td>\n",
       "      <td>6</td>\n",
       "      <td>UTF-8</td>\n",
       "      <td>Apache/2.4.10</td>\n",
       "      <td>15087.0</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>...</td>\n",
       "      <td>7</td>\n",
       "      <td>4</td>\n",
       "      <td>1230</td>\n",
       "      <td>17</td>\n",
       "      <td>19</td>\n",
       "      <td>1265</td>\n",
       "      <td>1230</td>\n",
       "      <td>17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B0_911</td>\n",
       "      <td>16</td>\n",
       "      <td>6</td>\n",
       "      <td>us-ascii</td>\n",
       "      <td>Microsoft-HTTPAPI/2.0</td>\n",
       "      <td>324.0</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B0_113</td>\n",
       "      <td>17</td>\n",
       "      <td>6</td>\n",
       "      <td>ISO-8859-1</td>\n",
       "      <td>nginx</td>\n",
       "      <td>162.0</td>\n",
       "      <td>US</td>\n",
       "      <td>AK</td>\n",
       "      <td>7/10/1997 4:00</td>\n",
       "      <td>12/09/2013 0:45</td>\n",
       "      <td>...</td>\n",
       "      <td>22</td>\n",
       "      <td>3</td>\n",
       "      <td>3812</td>\n",
       "      <td>39</td>\n",
       "      <td>37</td>\n",
       "      <td>18784</td>\n",
       "      <td>4380</td>\n",
       "      <td>39</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B0_403</td>\n",
       "      <td>17</td>\n",
       "      <td>6</td>\n",
       "      <td>UTF-8</td>\n",
       "      <td>None</td>\n",
       "      <td>124140.0</td>\n",
       "      <td>US</td>\n",
       "      <td>TX</td>\n",
       "      <td>12/05/1996 0:00</td>\n",
       "      <td>11/04/2017 0:00</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>4278</td>\n",
       "      <td>61</td>\n",
       "      <td>62</td>\n",
       "      <td>129889</td>\n",
       "      <td>4586</td>\n",
       "      <td>61</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       URL  URL_LENGTH  NUMBER_SPECIAL_CHARACTERS     CHARSET  \\\n",
       "0   M0_109          16                          7  iso-8859-1   \n",
       "1  B0_2314          16                          6       UTF-8   \n",
       "2   B0_911          16                          6    us-ascii   \n",
       "3   B0_113          17                          6  ISO-8859-1   \n",
       "4   B0_403          17                          6       UTF-8   \n",
       "\n",
       "                  SERVER  CONTENT_LENGTH WHOIS_COUNTRY WHOIS_STATEPRO  \\\n",
       "0                  nginx           263.0          None           None   \n",
       "1          Apache/2.4.10         15087.0          None           None   \n",
       "2  Microsoft-HTTPAPI/2.0           324.0          None           None   \n",
       "3                  nginx           162.0            US             AK   \n",
       "4                   None        124140.0            US             TX   \n",
       "\n",
       "      WHOIS_REGDATE WHOIS_UPDATED_DATE  ...   DIST_REMOTE_TCP_PORT  \\\n",
       "0  10/10/2015 18:21               None  ...                      0   \n",
       "1              None               None  ...                      7   \n",
       "2              None               None  ...                      0   \n",
       "3    7/10/1997 4:00    12/09/2013 0:45  ...                     22   \n",
       "4   12/05/1996 0:00    11/04/2017 0:00  ...                      2   \n",
       "\n",
       "   REMOTE_IPS  APP_BYTES  SOURCE_APP_PACKETS  REMOTE_APP_PACKETS  \\\n",
       "0           2        700                   9                  10   \n",
       "1           4       1230                  17                  19   \n",
       "2           0          0                   0                   0   \n",
       "3           3       3812                  39                  37   \n",
       "4           5       4278                  61                  62   \n",
       "\n",
       "   SOURCE_APP_BYTES  REMOTE_APP_BYTES  APP_PACKETS  DNS_QUERY_TIMES  Type  \n",
       "0              1153               832            9              2.0     1  \n",
       "1              1265              1230           17              0.0     0  \n",
       "2                 0                 0            0              0.0     0  \n",
       "3             18784              4380           39              8.0     0  \n",
       "4            129889              4586           61              4.0     0  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 217,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'URL': 0,\n",
       " 'CHARSET': 7,\n",
       " 'SERVER': 175,\n",
       " 'WHOIS_COUNTRY': 306,\n",
       " 'WHOIS_STATEPRO': 362,\n",
       " 'WHOIS_REGDATE': 127,\n",
       " 'WHOIS_UPDATED_DATE': 139}"
      ]
     },
     "execution_count": 218,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find string columns \n",
    "colStr = data.dtypes[data.dtypes== 'object']\n",
    "\n",
    "# count the 'None' for every columns\n",
    "colStr = data.dtypes[data.dtypes== 'object']\n",
    "calNone = {col: len(data[data.loc[:,col]=='None']) for col in colStr.index}\n",
    "calNone"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'URL': 1781,\n",
       " 'URL_LENGTH': 142,\n",
       " 'NUMBER_SPECIAL_CHARACTERS': 31,\n",
       " 'CHARSET': 9,\n",
       " 'SERVER': 240,\n",
       " 'CONTENT_LENGTH': 638,\n",
       " 'WHOIS_COUNTRY': 49,\n",
       " 'WHOIS_STATEPRO': 182,\n",
       " 'WHOIS_REGDATE': 891,\n",
       " 'WHOIS_UPDATED_DATE': 594,\n",
       " 'TCP_CONVERSATION_EXCHANGE': 103,\n",
       " 'DIST_REMOTE_TCP_PORT': 66,\n",
       " 'REMOTE_IPS': 18,\n",
       " 'APP_BYTES': 825,\n",
       " 'SOURCE_APP_PACKETS': 113,\n",
       " 'REMOTE_APP_PACKETS': 116,\n",
       " 'SOURCE_APP_BYTES': 885,\n",
       " 'REMOTE_APP_BYTES': 822,\n",
       " 'APP_PACKETS': 113,\n",
       " 'DNS_QUERY_TIMES': 11,\n",
       " 'Type': 2}"
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# count the unique for every columns\n",
    "calUni = {col:len(data.loc[:,col].unique()) for col in data.columns}\n",
    "calUni"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "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>URL</th>\n",
       "      <th>WHOIS_REGDATE</th>\n",
       "      <th>WHOIS_UPDATED_DATE</th>\n",
       "      <th>CONTENT_LENGTH</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>M0_109</td>\n",
       "      <td>10/10/2015 18:21</td>\n",
       "      <td>None</td>\n",
       "      <td>263.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B0_2314</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>15087.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B0_911</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>324.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B0_113</td>\n",
       "      <td>7/10/1997 4:00</td>\n",
       "      <td>12/09/2013 0:45</td>\n",
       "      <td>162.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B0_403</td>\n",
       "      <td>12/05/1996 0:00</td>\n",
       "      <td>11/04/2017 0:00</td>\n",
       "      <td>124140.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       URL     WHOIS_REGDATE WHOIS_UPDATED_DATE  CONTENT_LENGTH\n",
       "0   M0_109  10/10/2015 18:21               None           263.0\n",
       "1  B0_2314              None               None         15087.0\n",
       "2   B0_911              None               None           324.0\n",
       "3   B0_113    7/10/1997 4:00    12/09/2013 0:45           162.0\n",
       "4   B0_403   12/05/1996 0:00    11/04/2017 0:00        124140.0"
      ]
     },
     "execution_count": 220,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find columns that are totaally different and cannot used as the categoriable variables\n",
    "uniqueVal = ['URL','WHOIS_REGDATE','WHOIS_UPDATED_DATE','CONTENT_LENGTH']\n",
    "data.loc[:,uniqueVal].head() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>DNS_QUERY_TIMES</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SERVER</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>URL_LENGTH</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APP_BYTES</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APP_PACKETS</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0\n",
       "DNS_QUERY_TIMES  1\n",
       "SERVER           1\n",
       "URL_LENGTH       0\n",
       "APP_BYTES        0\n",
       "APP_PACKETS      0"
      ]
     },
     "execution_count": 221,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dropUniq = data.drop(columns=uniqueVal)\n",
    "nansum = pd.DataFrame(data_dropUniq.isna().sum()).sort_values(by=0, ascending = False)\n",
    "nansum.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1779"
      ]
     },
     "execution_count": 222,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# drop nan\n",
    "data_final_dropna = data_dropUniq.dropna() # drop all rows with nan\n",
    "len(data_final_dropna)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1779, 491)\n"
     ]
    },
    {
     "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>URL_LENGTH</th>\n",
       "      <th>NUMBER_SPECIAL_CHARACTERS</th>\n",
       "      <th>TCP_CONVERSATION_EXCHANGE</th>\n",
       "      <th>DIST_REMOTE_TCP_PORT</th>\n",
       "      <th>REMOTE_IPS</th>\n",
       "      <th>APP_BYTES</th>\n",
       "      <th>SOURCE_APP_PACKETS</th>\n",
       "      <th>REMOTE_APP_PACKETS</th>\n",
       "      <th>SOURCE_APP_BYTES</th>\n",
       "      <th>REMOTE_APP_BYTES</th>\n",
       "      <th>...</th>\n",
       "      <th>WHOIS_STATEPRO_kireka</th>\n",
       "      <th>WHOIS_STATEPRO_liaoningsheng</th>\n",
       "      <th>WHOIS_STATEPRO_nj</th>\n",
       "      <th>WHOIS_STATEPRO_ny</th>\n",
       "      <th>WHOIS_STATEPRO_qc</th>\n",
       "      <th>WHOIS_STATEPRO_quebec</th>\n",
       "      <th>WHOIS_STATEPRO_shandong</th>\n",
       "      <th>WHOIS_STATEPRO_va</th>\n",
       "      <th>WHOIS_STATEPRO_widestep@mail.ru</th>\n",
       "      <th>WHOIS_STATEPRO_worcs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>16</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>700</td>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>1153</td>\n",
       "      <td>832</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16</td>\n",
       "      <td>6</td>\n",
       "      <td>17</td>\n",
       "      <td>7</td>\n",
       "      <td>4</td>\n",
       "      <td>1230</td>\n",
       "      <td>17</td>\n",
       "      <td>19</td>\n",
       "      <td>1265</td>\n",
       "      <td>1230</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>16</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>17</td>\n",
       "      <td>6</td>\n",
       "      <td>31</td>\n",
       "      <td>22</td>\n",
       "      <td>3</td>\n",
       "      <td>3812</td>\n",
       "      <td>39</td>\n",
       "      <td>37</td>\n",
       "      <td>18784</td>\n",
       "      <td>4380</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17</td>\n",
       "      <td>6</td>\n",
       "      <td>57</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>4278</td>\n",
       "      <td>61</td>\n",
       "      <td>62</td>\n",
       "      <td>129889</td>\n",
       "      <td>4586</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 491 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   URL_LENGTH  NUMBER_SPECIAL_CHARACTERS  TCP_CONVERSATION_EXCHANGE  \\\n",
       "0          16                          7                          7   \n",
       "1          16                          6                         17   \n",
       "2          16                          6                          0   \n",
       "3          17                          6                         31   \n",
       "4          17                          6                         57   \n",
       "\n",
       "   DIST_REMOTE_TCP_PORT  REMOTE_IPS  APP_BYTES  SOURCE_APP_PACKETS  \\\n",
       "0                     0           2        700                   9   \n",
       "1                     7           4       1230                  17   \n",
       "2                     0           0          0                   0   \n",
       "3                    22           3       3812                  39   \n",
       "4                     2           5       4278                  61   \n",
       "\n",
       "   REMOTE_APP_PACKETS  SOURCE_APP_BYTES  REMOTE_APP_BYTES  \\\n",
       "0                  10              1153               832   \n",
       "1                  19              1265              1230   \n",
       "2                   0                 0                 0   \n",
       "3                  37             18784              4380   \n",
       "4                  62            129889              4586   \n",
       "\n",
       "           ...           WHOIS_STATEPRO_kireka  WHOIS_STATEPRO_liaoningsheng  \\\n",
       "0          ...                               0                             0   \n",
       "1          ...                               0                             0   \n",
       "2          ...                               0                             0   \n",
       "3          ...                               0                             0   \n",
       "4          ...                               0                             0   \n",
       "\n",
       "   WHOIS_STATEPRO_nj  WHOIS_STATEPRO_ny  WHOIS_STATEPRO_qc  \\\n",
       "0                  0                  0                  0   \n",
       "1                  0                  0                  0   \n",
       "2                  0                  0                  0   \n",
       "3                  0                  0                  0   \n",
       "4                  0                  0                  0   \n",
       "\n",
       "   WHOIS_STATEPRO_quebec  WHOIS_STATEPRO_shandong  WHOIS_STATEPRO_va  \\\n",
       "0                      0                        0                  0   \n",
       "1                      0                        0                  0   \n",
       "2                      0                        0                  0   \n",
       "3                      0                        0                  0   \n",
       "4                      0                        0                  0   \n",
       "\n",
       "   WHOIS_STATEPRO_widestep@mail.ru  WHOIS_STATEPRO_worcs  \n",
       "0                                0                     0  \n",
       "1                                0                     0  \n",
       "2                                0                     0  \n",
       "3                                0                     0  \n",
       "4                                0                     0  \n",
       "\n",
       "[5 rows x 491 columns]"
      ]
     },
     "execution_count": 223,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# tanfer dummies variables into list of different categories\n",
    "data_final_dum = pd.get_dummies(data_final_dropna)\n",
    "print(data_final_dum.shape)\n",
    "data_final_dum.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['WHOIS_STATEPRO_Greater London', 'WHOIS_STATEPRO_HANTS',\n",
       "       'WHOIS_STATEPRO_HR', 'WHOIS_STATEPRO_Haryana', 'WHOIS_STATEPRO_IL',\n",
       "       'WHOIS_STATEPRO_ILOCOS NORTE R3', 'WHOIS_STATEPRO_INDAL',\n",
       "       'WHOIS_STATEPRO_Illinois', 'WHOIS_STATEPRO_Indiana',\n",
       "       'WHOIS_STATEPRO_KG', 'WHOIS_STATEPRO_KS', 'WHOIS_STATEPRO_Kansas',\n",
       "       'WHOIS_STATEPRO_Karnataka', 'WHOIS_STATEPRO_Kentucky',\n",
       "       'WHOIS_STATEPRO_Krasnoyarsk', 'WHOIS_STATEPRO_LA',\n",
       "       'WHOIS_STATEPRO_London', 'WHOIS_STATEPRO_MA',\n",
       "       'WHOIS_STATEPRO_MAHARASHTR', 'WHOIS_STATEPRO_MAINE',\n",
       "       'WHOIS_STATEPRO_MB', 'WHOIS_STATEPRO_MD', 'WHOIS_STATEPRO_ME',\n",
       "       'WHOIS_STATEPRO_MH', 'WHOIS_STATEPRO_MI', 'WHOIS_STATEPRO_MIDDLESEX',\n",
       "       'WHOIS_STATEPRO_MO', 'WHOIS_STATEPRO_MOSCOW',\n",
       "       'WHOIS_STATEPRO_Maharashtra', 'WHOIS_STATEPRO_Mahe',\n",
       "       'WHOIS_STATEPRO_Maine', 'WHOIS_STATEPRO_Manila',\n",
       "       'WHOIS_STATEPRO_Manitoba', 'WHOIS_STATEPRO_Maryland',\n",
       "       'WHOIS_STATEPRO_Massachusetts', 'WHOIS_STATEPRO_Metro Manila',\n",
       "       'WHOIS_STATEPRO_Michigan', 'WHOIS_STATEPRO_Minnesota',\n",
       "       'WHOIS_STATEPRO_Missouri', 'WHOIS_STATEPRO_Montevideo',\n",
       "       'WHOIS_STATEPRO_NC', 'WHOIS_STATEPRO_NEW SOUTH WALES',\n",
       "       'WHOIS_STATEPRO_NEW YORK', 'WHOIS_STATEPRO_NH', 'WHOIS_STATEPRO_NJ',\n",
       "       'WHOIS_STATEPRO_NONE', 'WHOIS_STATEPRO_NOT APPLICABLE',\n",
       "       'WHOIS_STATEPRO_NSW', 'WHOIS_STATEPRO_NV', 'WHOIS_STATEPRO_NY'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 224,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "listcol = data_final_dum.columns\n",
    "listcol[350:400]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Build both Random Forest (10 points) and Naive Bayes (10 points) classifiers to predict the TYPE variable.  Be sure to provide interpretations of your results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_web_train,df_web_test = train_test_split(data_final_dum, test_size=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda2/envs/mypython/lib/python3.7/site-packages/ipykernel_launcher.py:15: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  from ipykernel import kernelapp as app\n"
     ]
    }
   ],
   "source": [
    "# build a random forest\n",
    "\n",
    "df_web_train_dropY = df_web_train.loc[:, df_web_train.columns != 'Type']\n",
    "df_web_test_dropY = df_web_test.loc[:, df_web_test.columns != 'Type']\n",
    "\n",
    "# df_web_train_dropY = df_web_train.iloc[:,:len(df_web_train.columns)-1]\n",
    "# df_web_test_dropY = df_web_test.iloc[:,:len(df_web_train.columns)-1]\n",
    "\n",
    "rf_model = skens.RandomForestClassifier(n_estimators=10, oob_score=False, criterion='entropy')\n",
    "# oob_score: whether to use out-of-bag samples to estimate the R^2 on unseen data.\n",
    "rf_model.fit(df_web_train_dropY,df_web_train.Type)\n",
    "\n",
    "# this is testing the model \n",
    "predicted_labels = rf_model.predict(df_web_test_dropY)\n",
    "df_web_test['predicted_rf_tree'] = predicted_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x108316748>"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x1440 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check feature importance with biggest 20 features\n",
    "feat_importance = rf_model.feature_importances_\n",
    "biggest = pd.DataFrame({'Feature Importance':feat_importance},index=df_web_test_dropY.columns).sort_values(by='Feature Importance', ascending=False)\n",
    "biggest[:20].plot(kind='barh', figsize=(30,20))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.8875502]\n"
     ]
    }
   ],
   "source": [
    "# # get model score for RF\n",
    "# rf_model.fit(df_web_train_dropY,df_web_train.Type)\n",
    "# print(rf_model.score(df_web_test_dropY, df_web_test.Type))\n",
    "\n",
    "# anotehr way : using cv， assumed parameters\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "params = {\n",
    "                 'n_estimators': [5],\n",
    "                 'max_depth': [2],\n",
    "            }\n",
    "RF_CV = GridSearchCV(rf_model, cv=10, param_grid=params, iid = True)\n",
    "RF_CV.fit(df_web_train_dropY,df_web_train.Type)\n",
    "print(RF_CV.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.2 Naive Bayes (10 points) classifiers to predict the TYPE variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GaussianNB(priors=None, var_smoothing=1e-09)"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# train the model by Naive Bayes\n",
    "gnb_model = sknb.GaussianNB()\n",
    "# given sepal length, predict if setosa\n",
    "gnb_model.fit(df_web_train_dropY,df_web_train.Type)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [],
   "source": [
    "# test the model\n",
    "df_web_testNB = df_web_test.copy()\n",
    "y_pred = gnb_model.predict(df_web_test_dropY)\n",
    "df_web_testNB['predicted_nb'] = y_pred\n",
    "# comparePlot(df_web_testNB,\"Type\",\"predicted_nb\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.24899598]\n"
     ]
    }
   ],
   "source": [
    "# get model score for NB\n",
    "# gnb_model.fit(df_web_train_dropY,df_web_train.Type)\n",
    "# print(gnb_model.score(df_web_test_dropY, df_web_test.Type))\n",
    "\n",
    "# another way\n",
    "# cal cv errors for NB \n",
    "params = {}\n",
    "NB_CV = GridSearchCV(gnb_model, cv=10, param_grid=params, iid = True)\n",
    "NB_CV.fit(df_web_train_dropY,df_web_train.Type)\n",
    "print(NB_CV.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As the column \"CONTENT_LENGTH\" has too much NaN, it is dropped and dropna was used for dropping nan in different rows. From the Random Forest, the top 20 influential features was plotted above. Random Forest model was built and its CV erros with 10 folders was about 88% accuracy, but Naive Bayes only has 38% accuracy. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Experiment with changing the number of \"folds\" in your cross-validator for your Random Forest classifiers (10 points). Be sure to provide interpretations of your results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=10, error_score='raise-deprecating',\n",
       "       estimator=RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
       "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None, param_grid={},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring=None, verbose=0)"
      ]
     },
     "execution_count": 252,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Model Selection and Comparison\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "param_grid = {} # fixex param for all cv situations\n",
    "cv_grid = [10, 100, 200]\n",
    "\n",
    "grid_clf = GridSearchCV(rf_model, param_grid, cv=cv_grid[0])\n",
    "grid_clf.fit(df_web_train_dropY,df_web_train.Type)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
       "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False)"
      ]
     },
     "execution_count": 253,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_clf.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.95180723])"
      ]
     },
     "execution_count": 254,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_clf.cv_results_['mean_test_score']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
      "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "            min_samples_leaf=1, min_samples_split=2,\n",
      "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
      "            oob_score=False, random_state=None, verbose=0,\n",
      "            warm_start=False)\n",
      "[0.95582329]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda2/envs/mypython/lib/python3.7/site-packages/sklearn/model_selection/_search.py:841: DeprecationWarning: The default of the `iid` parameter will change from True to False in version 0.22 and will be removed in 0.24. This will change numeric results when test-set sizes are unequal.\n",
      "  DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "# for another cv condition cv = 100\n",
    "grid_clf = GridSearchCV(rf_model, param_grid, cv=cv_grid[1])\n",
    "grid_clf.fit(df_web_train_dropY,df_web_train.Type)\n",
    "print(grid_clf.best_estimator_)\n",
    "print(grid_clf.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda2/envs/mypython/lib/python3.7/site-packages/sklearn/model_selection/_split.py:652: Warning: The least populated class in y has only 148 members, which is too few. The minimum number of members in any class cannot be less than n_splits=200.\n",
      "  % (min_groups, self.n_splits)), Warning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
      "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "            min_samples_leaf=1, min_samples_split=2,\n",
      "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
      "            oob_score=False, random_state=None, verbose=0,\n",
      "            warm_start=False)\n",
      "[0.95742972]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda2/envs/mypython/lib/python3.7/site-packages/sklearn/model_selection/_search.py:841: DeprecationWarning: The default of the `iid` parameter will change from True to False in version 0.22 and will be removed in 0.24. This will change numeric results when test-set sizes are unequal.\n",
      "  DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "# for another cv condition cv = 15\n",
    "grid_clf = GridSearchCV(rf_model, param_grid, cv=cv_grid[2])\n",
    "grid_clf.fit(df_web_train_dropY,df_web_train.Type)\n",
    "print(grid_clf.best_estimator_)\n",
    "print(grid_clf.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fixed Parameters, with different value of folds for CV methdo, the accuracies incrase a little, while still are closed to each other, "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Tune your Random Forest model by varying hyperparamters (15 points). Be sure to provide interpretations of your results.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=10, error_score='raise-deprecating',\n",
       "       estimator=RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
       "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'n_estimators': [5, 10, 15, 20, 25], 'max_depth': [2, 5, 7, 9]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring=None, verbose=0)"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Model Selection and Comparison\n",
    "param_grid = {\n",
    "                 'n_estimators': [5, 10, 15, 20, 25],\n",
    "                 'max_depth': [2, 5, 7, 9],\n",
    "             }\n",
    "grid_clf = GridSearchCV(rf_model, param_grid, cv=10)\n",
    "grid_clf.fit(df_web_train_dropY,df_web_train.Type)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'max_depth': 9, 'n_estimators': 5}"
      ]
     },
     "execution_count": 238,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_clf.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
       "            max_depth=9, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=5, n_jobs=None,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False)"
      ]
     },
     "execution_count": 239,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_clf.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.89076305, 0.88353414, 0.88514056, 0.88273092, 0.8811245 ,\n",
       "       0.90763052, 0.91084337, 0.90843373, 0.91084337, 0.91164659,\n",
       "       0.92048193, 0.91967871, 0.9188755 , 0.91566265, 0.91646586,\n",
       "       0.93815261, 0.93413655, 0.93493976, 0.92851406, 0.93413655])"
      ]
     },
     "execution_count": 240,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_clf.cv_results_['mean_test_score']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9381526104417671"
      ]
     },
     "execution_count": 241,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_clf.cv_results_['mean_test_score'].max()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With differet hyperparamters: n_estimators, max_depth, the best model has parameters {'max_depth': 9, 'n_estimators': 10}, and its highest average accuracy is 93.97% under 10 folder CV."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Experiment with setting the appropriate values for priors in the Naive Bayes model (the priors should correspond to the proportions of the TYPE values) (10 points). Be sure to provide interpretations of your results.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.8811245, 0.1188755])"
      ]
     },
     "execution_count": 242,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gnb_model.class_prior_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8811244979919679"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# cal the percent of zero in train dataset\n",
    "len(df_web_train[df_web_train.Type == 0])/len(df_web_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {},
   "outputs": [],
   "source": [
    "# assume prior is half half\n",
    "gnb_model_newPr = sknb.GaussianNB(priors=[0.5, 0.5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.23614458]\n"
     ]
    }
   ],
   "source": [
    "# cal cv errors for NB \n",
    "params = {}\n",
    "NB_CV_new = GridSearchCV(gnb_model_newPr, cv=10, param_grid=params, iid = True) # \n",
    "NB_CV_new.fit(df_web_train_dropY,df_web_train.Type)\n",
    "print(NB_CV_new.cv_results_['mean_test_score'])\n",
    "\n",
    "# another way:\n",
    "# get model score\n",
    "# gnb_model_newPr.fit(df_web_train_dropY,df_web_train.Type)\n",
    "# gnb_model_newPr.score(df_web_test_dropY, df_web_test.Type)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the calculateion of zero in TYPE, we foundt the original NB model uses the prior in dataset. If we assume both 0 and 1 has 50% prior, with same CV method, accuracy of NB method is reduved from 38% to 35%, supporting the importance of suitable prior choice."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5. Report the most important feature(s) based on your Random Forest classifier (10 points). Are you surprised by the results? Explain why.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from q2, we find the optimal RF model\n",
    "OptRF_model = grid_clf.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a1e378b00>"
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x1440 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check feature importance with biggest 20 features\n",
    "feat_importance = OptRF_model.feature_importances_\n",
    "biggest_opt = pd.DataFrame({'Feature Importance':feat_importance},index=df_web_test_dropY.columns).sort_values(by='Feature Importance', ascending=False)\n",
    "biggest_opt[:20].plot(kind='barh', figsize=(30,20))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "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>index</th>\n",
       "      <th>Feature Importance_x</th>\n",
       "      <th>Feature Importance_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NUMBER_SPECIAL_CHARACTERS</td>\n",
       "      <td>0.084973</td>\n",
       "      <td>0.032292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>WHOIS_COUNTRY_ES</td>\n",
       "      <td>0.071669</td>\n",
       "      <td>0.089280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>SOURCE_APP_BYTES</td>\n",
       "      <td>0.063098</td>\n",
       "      <td>0.091900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>APP_BYTES</td>\n",
       "      <td>0.056615</td>\n",
       "      <td>0.019521</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>DIST_REMOTE_TCP_PORT</td>\n",
       "      <td>0.056325</td>\n",
       "      <td>0.024927</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>URL_LENGTH</td>\n",
       "      <td>0.054701</td>\n",
       "      <td>0.032121</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>REMOTE_APP_BYTES</td>\n",
       "      <td>0.052599</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>WHOIS_COUNTRY_US</td>\n",
       "      <td>0.052410</td>\n",
       "      <td>0.026756</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>REMOTE_APP_PACKETS</td>\n",
       "      <td>0.045925</td>\n",
       "      <td>0.110138</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>REMOTE_IPS</td>\n",
       "      <td>0.040611</td>\n",
       "      <td>0.069131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>SOURCE_APP_PACKETS</td>\n",
       "      <td>0.040282</td>\n",
       "      <td>0.021775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>TCP_CONVERSATION_EXCHANGE</td>\n",
       "      <td>0.040155</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>DNS_QUERY_TIMES</td>\n",
       "      <td>0.029717</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>APP_PACKETS</td>\n",
       "      <td>0.026487</td>\n",
       "      <td>0.044263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>CHARSET_UTF-8</td>\n",
       "      <td>0.022461</td>\n",
       "      <td>0.019058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>WHOIS_COUNTRY_None</td>\n",
       "      <td>0.022369</td>\n",
       "      <td>0.021538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>SERVER_nginx</td>\n",
       "      <td>0.021407</td>\n",
       "      <td>0.018939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>WHOIS_STATEPRO_Utah</td>\n",
       "      <td>0.016452</td>\n",
       "      <td>0.020611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>SERVER_None</td>\n",
       "      <td>0.014118</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>WHOIS_STATEPRO_None</td>\n",
       "      <td>0.013316</td>\n",
       "      <td>0.038092</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>WHOIS_STATEPRO_Barcelona</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.079081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>CHARSET_utf-8</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.028914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>SERVER_Microsoft-IIS/6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.021410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>WHOIS_STATEPRO_CA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.018666</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        index  Feature Importance_x  Feature Importance_y\n",
       "0   NUMBER_SPECIAL_CHARACTERS              0.084973              0.032292\n",
       "1            WHOIS_COUNTRY_ES              0.071669              0.089280\n",
       "2            SOURCE_APP_BYTES              0.063098              0.091900\n",
       "3                   APP_BYTES              0.056615              0.019521\n",
       "4        DIST_REMOTE_TCP_PORT              0.056325              0.024927\n",
       "5                  URL_LENGTH              0.054701              0.032121\n",
       "6            REMOTE_APP_BYTES              0.052599                   NaN\n",
       "7            WHOIS_COUNTRY_US              0.052410              0.026756\n",
       "8          REMOTE_APP_PACKETS              0.045925              0.110138\n",
       "9                  REMOTE_IPS              0.040611              0.069131\n",
       "10         SOURCE_APP_PACKETS              0.040282              0.021775\n",
       "11  TCP_CONVERSATION_EXCHANGE              0.040155                   NaN\n",
       "12            DNS_QUERY_TIMES              0.029717                   NaN\n",
       "13                APP_PACKETS              0.026487              0.044263\n",
       "14              CHARSET_UTF-8              0.022461              0.019058\n",
       "15         WHOIS_COUNTRY_None              0.022369              0.021538\n",
       "16               SERVER_nginx              0.021407              0.018939\n",
       "17        WHOIS_STATEPRO_Utah              0.016452              0.020611\n",
       "18                SERVER_None              0.014118                   NaN\n",
       "19        WHOIS_STATEPRO_None              0.013316              0.038092\n",
       "20   WHOIS_STATEPRO_Barcelona                   NaN              0.079081\n",
       "21              CHARSET_utf-8                   NaN              0.028914\n",
       "22   SERVER_Microsoft-IIS/6.0                   NaN              0.021410\n",
       "23          WHOIS_STATEPRO_CA                   NaN              0.018666"
      ]
     },
     "execution_count": 248,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# biggest is sorted features from original rf model; biggest_opt is sorted features from new model\n",
    "pd.merge(biggest[:20].reset_index(), biggest_opt[:20].reset_index(), on='index', how= 'outer')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It seems there is not much difference between top 20 features from the original RF and optimal RF, as there are only 3 new important top 20 features. Surprising thing for me is that the importantce of same features will change, such as 'REMOTE_APP_PACKETS' increased form 8% to 10%, while 'TCP_CONVERSATION_EXCHANGE' drops from 7% to 3%"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Summarize your findings, including which classifier and paramters gave you the best results and include an in-depth explanation of how you might improve your results. You should also investigate what the different fields mean, so you'll need to do some research about websites and internet protocols! (15 points)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "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>URL_LENGTH</th>\n",
       "      <th>NUMBER_SPECIAL_CHARACTERS</th>\n",
       "      <th>CONTENT_LENGTH</th>\n",
       "      <th>TCP_CONVERSATION_EXCHANGE</th>\n",
       "      <th>DIST_REMOTE_TCP_PORT</th>\n",
       "      <th>REMOTE_IPS</th>\n",
       "      <th>APP_BYTES</th>\n",
       "      <th>SOURCE_APP_PACKETS</th>\n",
       "      <th>REMOTE_APP_PACKETS</th>\n",
       "      <th>SOURCE_APP_BYTES</th>\n",
       "      <th>REMOTE_APP_BYTES</th>\n",
       "      <th>APP_PACKETS</th>\n",
       "      <th>DNS_QUERY_TIMES</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>969.000000</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1.781000e+03</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1.781000e+03</td>\n",
       "      <td>1.781000e+03</td>\n",
       "      <td>1781.000000</td>\n",
       "      <td>1780.000000</td>\n",
       "      <td>1781.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>56.961258</td>\n",
       "      <td>11.111735</td>\n",
       "      <td>11726.927761</td>\n",
       "      <td>16.261089</td>\n",
       "      <td>5.472768</td>\n",
       "      <td>3.060640</td>\n",
       "      <td>2.982339e+03</td>\n",
       "      <td>18.540146</td>\n",
       "      <td>18.746210</td>\n",
       "      <td>1.589255e+04</td>\n",
       "      <td>3.155599e+03</td>\n",
       "      <td>18.540146</td>\n",
       "      <td>2.263483</td>\n",
       "      <td>0.121280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>27.555586</td>\n",
       "      <td>4.549896</td>\n",
       "      <td>36391.809051</td>\n",
       "      <td>40.500975</td>\n",
       "      <td>21.807327</td>\n",
       "      <td>3.386975</td>\n",
       "      <td>5.605057e+04</td>\n",
       "      <td>41.627173</td>\n",
       "      <td>46.397969</td>\n",
       "      <td>6.986193e+04</td>\n",
       "      <td>5.605378e+04</td>\n",
       "      <td>41.627173</td>\n",
       "      <td>2.930853</td>\n",
       "      <td>0.326544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>16.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>39.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>324.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>49.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>1853.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>6.720000e+02</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>5.790000e+02</td>\n",
       "      <td>7.350000e+02</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>68.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>11323.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>2.328000e+03</td>\n",
       "      <td>26.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>9.806000e+03</td>\n",
       "      <td>2.701000e+03</td>\n",
       "      <td>26.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>249.000000</td>\n",
       "      <td>43.000000</td>\n",
       "      <td>649263.000000</td>\n",
       "      <td>1194.000000</td>\n",
       "      <td>708.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>2.362906e+06</td>\n",
       "      <td>1198.000000</td>\n",
       "      <td>1284.000000</td>\n",
       "      <td>2.060012e+06</td>\n",
       "      <td>2.362906e+06</td>\n",
       "      <td>1198.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        URL_LENGTH  NUMBER_SPECIAL_CHARACTERS  CONTENT_LENGTH  \\\n",
       "count  1781.000000                1781.000000      969.000000   \n",
       "mean     56.961258                  11.111735    11726.927761   \n",
       "std      27.555586                   4.549896    36391.809051   \n",
       "min      16.000000                   5.000000        0.000000   \n",
       "25%      39.000000                   8.000000      324.000000   \n",
       "50%      49.000000                  10.000000     1853.000000   \n",
       "75%      68.000000                  13.000000    11323.000000   \n",
       "max     249.000000                  43.000000   649263.000000   \n",
       "\n",
       "       TCP_CONVERSATION_EXCHANGE  DIST_REMOTE_TCP_PORT   REMOTE_IPS  \\\n",
       "count                1781.000000           1781.000000  1781.000000   \n",
       "mean                   16.261089              5.472768     3.060640   \n",
       "std                    40.500975             21.807327     3.386975   \n",
       "min                     0.000000              0.000000     0.000000   \n",
       "25%                     0.000000              0.000000     0.000000   \n",
       "50%                     7.000000              0.000000     2.000000   \n",
       "75%                    22.000000              5.000000     5.000000   \n",
       "max                  1194.000000            708.000000    17.000000   \n",
       "\n",
       "          APP_BYTES  SOURCE_APP_PACKETS  REMOTE_APP_PACKETS  SOURCE_APP_BYTES  \\\n",
       "count  1.781000e+03         1781.000000         1781.000000      1.781000e+03   \n",
       "mean   2.982339e+03           18.540146           18.746210      1.589255e+04   \n",
       "std    5.605057e+04           41.627173           46.397969      6.986193e+04   \n",
       "min    0.000000e+00            0.000000            0.000000      0.000000e+00   \n",
       "25%    0.000000e+00            0.000000            0.000000      0.000000e+00   \n",
       "50%    6.720000e+02            8.000000            9.000000      5.790000e+02   \n",
       "75%    2.328000e+03           26.000000           25.000000      9.806000e+03   \n",
       "max    2.362906e+06         1198.000000         1284.000000      2.060012e+06   \n",
       "\n",
       "       REMOTE_APP_BYTES  APP_PACKETS  DNS_QUERY_TIMES         Type  \n",
       "count      1.781000e+03  1781.000000      1780.000000  1781.000000  \n",
       "mean       3.155599e+03    18.540146         2.263483     0.121280  \n",
       "std        5.605378e+04    41.627173         2.930853     0.326544  \n",
       "min        0.000000e+00     0.000000         0.000000     0.000000  \n",
       "25%        0.000000e+00     0.000000         0.000000     0.000000  \n",
       "50%        7.350000e+02     8.000000         0.000000     0.000000  \n",
       "75%        2.701000e+03    26.000000         4.000000     0.000000  \n",
       "max        2.362906e+06  1198.000000        20.000000     1.000000  "
      ]
     },
     "execution_count": 249,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The best model is the has parameters {'max_depth': 9, 'n_estimators': 10}, and its average accuracy is 93.97%. \n",
    "\n",
    "Some important feasuters explanation: \n",
    "REMOTE_IPS: this variable has the total number of IPs connected to the honeypot\n",
    "SOURCE_APP_PACKETS: packets sent from the honeypot to the server\n",
    "\n",
    "So far, we drop several columns with too much differed values, such as columsn:'URL','WHOIS_REGDATE','WHOIS_UPDATED_DATE','CONTENT_LENGTH'. But, for example, WHOIS_REGDATE: Whois provides the server registration date, 'CONTENT_LENGTH' it represents the content size of the HTTP header, which can influence the Type ( 1 is for malicious websites and 0 is for benign websites). \n",
    "\n",
    "Therefore, we might divided 'WHOIS_REGDATE' specific dates into years, and categorize the length by several intervasl, such as 0-10000, and so on and so forth. Then, these information can be used for prediction and without too many number of factor levels. "
   ]
  },
  {
   "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
}