Phase 5 - Machine Learning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ab86e487",
   "metadata": {},
   "source": [
    "# Data Science Project | Common Vulnerability Exposures \n",
    "# Phase 5 - Machine Learning\n",
    "\n",
    "### During this phase we created 5 functions:\n",
    "\n",
    "1. **function load_dataset** - will split the dataframe into a new dataframe {x} which will contain the features vactor and a Series {y} which will contain our target values.\n",
    "2. **function split_to_train_and_test** - will will split the 'X' dataframe into 'X_train' and 'X_test', where the ratio of the test out of 'X' is 0.3 and the random state is 41. The 'y' series is splitted in a corresponding way into 'y_train' and 'y_test'.\n",
    "3. **function train_model** - will train a model, which will predict the CVE Score of the given vulnerabilities using linear regression.\n",
    "4. **function predict** - will predict the CVE Score for each of the vulnerabilities in the test set using the trained model. \n",
    "4. **function evaluate_performance** - will evaluate the performance of the model on the test set using R2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "18103436",
   "metadata": {},
   "source": [
    "##### Phase 3 functions and importing all libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "6e935644",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import all libraries\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "import sklearn\n",
    "from sklearn import linear_model\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "def load_dataframe(filename):\n",
    "    return pd.read_csv(filename)\n",
    "def remove_unnecessary_data(dataframe):\n",
    "    \n",
    "    df = dataframe.copy()\n",
    "    \n",
    "    #remove a column that contain only None\n",
    "    df.drop('Gained Access Level', inplace=True, axis=1)\n",
    "    \n",
    "    #remove all the rows with vulnerability score 0\n",
    "    df = df[df.Score != 0.0]\n",
    "    \n",
    "    #remove all the vulnerabilities from over a decade\n",
    "    for i in range(1998,2012):\n",
    "        df = df[df[\"CVE_ID\"].str.contains(str(i)) == False]\n",
    "    \n",
    "    #remove duplicate rows\n",
    "    df = df.drop_duplicates(keep='first').copy() \n",
    "    \n",
    "    return df\n",
    "def replace_missing_data(dataframe):\n",
    "    # Fill null values with 0\n",
    "    dt = dataframe.copy()\n",
    "    dt['Authentication'].mask(dt['Authentication'] == '???', 0, inplace=True)\n",
    "    dt = dt.fillna(0)\n",
    "    \n",
    "    return dt\n",
    "def convert_string_to_int(dataframe):\n",
    "    \n",
    "    df = dataframe.copy()\n",
    "        \n",
    "    # Convert CVE_ID from string to int\n",
    "    df['CVE_ID'] = df['CVE_ID'].str.replace(\"-\",\"\")\n",
    "    df['CVE_ID'] = df['CVE_ID'].str.replace(\"CVE\",\"\")\n",
    "    \n",
    "    # Convert Vulnerability Type to unique values\n",
    "    df.rename(columns = {'Vulnerability Type': 'VulnerabilityType'}, inplace=True)\n",
    "    le = LabelEncoder()\n",
    "    df['VulnerabilityType'] = le.fit_transform(df.VulnerabilityType)\n",
    "    \n",
    "    # Convert date to datetime \n",
    "    df['Publish Date'] = pd.to_datetime(df['Publish Date'])\n",
    "    df['Update Date'] = pd.to_datetime(df['Update Date'])\n",
    "    \n",
    "    # Extract a year from the date time\n",
    "    df['Update Date'] = pd.DatetimeIndex(df['Update Date']).year\n",
    "    df['Publish Date'] = pd.DatetimeIndex(df['Publish Date']).year\n",
    "    \n",
    "    # Convert Access\n",
    "    replaceAccess = {'Local':1, 'Local Network':2, 'Remote':3}\n",
    "    df.replace(replaceAccess,inplace=True)\n",
    "    \n",
    "    # Convert Complexity\n",
    "    replaceComplexity = {'Low':1, 'Medium':2, 'High':3}\n",
    "    df.replace(replaceComplexity,inplace=True)\n",
    "    \n",
    "    # Convert Authentication\n",
    "    replaceAuthentication = {'Not required':1, 'Single system':2}\n",
    "    df.replace(replaceAuthentication,inplace=True)\n",
    "    \n",
    "    # Convert Confidentiality\n",
    "    replaceConfidentiality = {'None':1, 'Partial':2, 'Complete':3}\n",
    "    df.replace(replaceConfidentiality,inplace=True)\n",
    "    \n",
    "    # Convert Integrity\n",
    "    replaceIntegrity = {'None':1, 'Partial':2, 'Complete':3}\n",
    "    df.replace(replaceIntegrity,inplace=True)\n",
    "    \n",
    "    # Convert Availability\n",
    "    replaceAvailability = {'None':1, 'Partial':2, 'Complete':3}\n",
    "    df.replace(replaceAvailability,inplace=True)\n",
    "\n",
    "    return df\n",
    "df = load_dataframe('cve_DF.csv')\n",
    "df = remove_unnecessary_data(df)\n",
    "dt = replace_missing_data(df)\n",
    "ds = convert_string_to_int(dt)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a40e1364",
   "metadata": {},
   "source": [
    "##### Phase 4 functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f071e9dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "#This function will split the dataframe into a new dataframe {x} which will contain the features vactor\n",
    "#and a Series {y} which will contain our target values\n",
    "def load_dataset(dataframe):\n",
    "    ds = dataframe.copy()\n",
    "    TRAINING_FEATURES = ds.columns[ds.columns != 'Score']\n",
    "    TARGET_FEATURE = 'Score'\n",
    "    X = ds[TRAINING_FEATURES]\n",
    "    y = ds[TARGET_FEATURE]\n",
    "    return X,y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "a307b030",
   "metadata": {},
   "outputs": [],
   "source": [
    "#This function will split the 'X' dataframe into 'X_train' and 'X_test', where the ratio of the test out of 'X' is 0.3\n",
    "#and the random state is 41.\n",
    "#The 'y' series is splitted in a corresponding way into 'y_train' and 'y_test'.\n",
    "def split_to_train_and_test(X, y):\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=41)\n",
    "    return X_train, X_test, y_train, y_test\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "9c5efb1c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#This function will train a model, which will predict the CVE Score of the given vulnerabilities using linear regression.\n",
    "def train_model(X_train, y_train):\n",
    "    lr = linear_model.LinearRegression()\n",
    "    return lr.fit(X=X_train, y=y_train);\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "bc7af5bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "#This function will predict the CVE Score for each of the vulnerabilities in the test set using the trained model.\n",
    "def predict(trained_model, X_test):\n",
    "    return trained_model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6f1f0410",
   "metadata": {},
   "outputs": [],
   "source": [
    "#This function will evaluate the performance of the model on the test set using R2.\n",
    "def evaluate_performance(y_test,y_predicted):\n",
    "    return r2_score(y_test,y_predicted)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c695e5ff",
   "metadata": {},
   "source": [
    "##### Running the functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "4cbb9328",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial amount of samples: 123144\n",
      "\n",
      "Number of training samples:  86200\n",
      "\n",
      "Number of test samples:  36944\n"
     ]
    }
   ],
   "source": [
    "X,y = load_dataset(ds) #split the dataframe\n",
    "X_train, X_test, y_train, y_test = split_to_train_and_test(X, y) #split into train and test\n",
    "\n",
    "print(\"Initial amount of samples:\",X.shape[0])\n",
    "print(\"\\nNumber of training samples: \",X_train.shape[0])\n",
    "print(\"\\nNumber of test samples: \", X_test.shape[0])\n",
    "\n",
    "trained_model = train_model(X_train, y_train) #train a model\n",
    "pred_vals = predict(trained_model, X_test) #predict values on the test set\n",
    "y_pred= pd.Series(pred_vals,index=X_test.index) #insert the values into a pandas series\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "553dbd87",
   "metadata": {},
   "source": [
    "##### Evaluate performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "97efb9b1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The evaluation score is: 0.9595019848744419\n",
      "\n",
      "Number of similar values out of 36944 values are:  33909\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>Actual</th>\n",
       "      <th>Predicted</th>\n",
       "      <th>correct</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>130612</th>\n",
       "      <td>7.2</td>\n",
       "      <td>7.517789</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163645</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.209840</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82418</th>\n",
       "      <td>5.5</td>\n",
       "      <td>5.142920</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75040</th>\n",
       "      <td>5.0</td>\n",
       "      <td>5.105596</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164510</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.243874</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132762</th>\n",
       "      <td>7.8</td>\n",
       "      <td>6.778744</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44997</th>\n",
       "      <td>4.0</td>\n",
       "      <td>4.259395</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63935</th>\n",
       "      <td>5.0</td>\n",
       "      <td>5.485610</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134361</th>\n",
       "      <td>7.8</td>\n",
       "      <td>6.773881</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101382</th>\n",
       "      <td>6.5</td>\n",
       "      <td>6.332927</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>36944 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Actual  Predicted  correct\n",
       "130612     7.2   7.517789        0\n",
       "163645    10.0  10.209840        0\n",
       "82418      5.5   5.142920        0\n",
       "75040      5.0   5.105596        0\n",
       "164510    10.0  10.243874        0\n",
       "...        ...        ...      ...\n",
       "132762     7.8   6.778744        1\n",
       "44997      4.0   4.259395        0\n",
       "63935      5.0   5.485610        0\n",
       "134361     7.8   6.773881        1\n",
       "101382     6.5   6.332927        0\n",
       "\n",
       "[36944 rows x 3 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eval_res = evaluate_performance(y_test, y_pred) #evaluate the performance of the model\n",
    "print(\"\\nThe evaluation score is:\", eval_res) \n",
    "\n",
    "#Presenting the actual CVE Score comparing to the Predicted score in a dataframe.\n",
    "#The correct column indicates weather the absolute worth of the predicted value is the same as the absolute worth of\n",
    "#the Actual value, 0 inidcates that the values are the same in absolute worth.\n",
    "\n",
    "resDF=pd.DataFrame({\"Actual\":y_test,\"Predicted\":y_pred})\n",
    "resDF[\"correct\"]=resDF[\"Actual\"].astype(int)-resDF[\"Predicted\"].astype(int)\n",
    "count = (resDF['correct'] == 0).sum()\n",
    "print(\"\\nNumber of similar values out of 36944 values are: \",count)\n",
    "resDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "991ba251",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}