{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data.csv',error_bad_lines = False, warn_bad_lines=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>password</th>\n",
       "      <th>strength</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>kzde5577</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>kino3434</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>visi7k1yr</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>megzy123</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>lamborghin1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      password  strength\n",
       "0     kzde5577         1\n",
       "1     kino3434         1\n",
       "2    visi7k1yr         1\n",
       "3     megzy123         1\n",
       "4  lamborghin1         1"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x22830ab7668>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(df['strength'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['char_count'] = df['password'].str.len() #Number of characters in password\n",
    "df['numerics'] = df['password'].apply(lambda x: len([str(x) for x in str(x) if x.isdigit()])) #No. of numerals in password\n",
    "df['alpha'] = df['password'].apply(lambda x: len([x for x in str(x) if x.isalpha()])) #No. of alphabets in password\n",
    "\n",
    "vowels = ['a', 'e', 'i', 'o', 'u']\n",
    "#To check if the password is an actual english word or name, I extracted the number of vowels and consonants in the text\n",
    "#The basic idea is: The more the consonants, less meaning the password makes\n",
    "df['vowels'] = df['password'].apply(lambda x: len([x for x in str(x) if x in vowels]))\n",
    "df['consonants'] = df['password'].apply(lambda x: len([x for x in str(x) if x not in vowels and x.isalpha()]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>password</th>\n",
       "      <th>strength</th>\n",
       "      <th>char_count</th>\n",
       "      <th>numerics</th>\n",
       "      <th>alpha</th>\n",
       "      <th>vowels</th>\n",
       "      <th>consonants</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>kzde5577</td>\n",
       "      <td>1</td>\n",
       "      <td>8.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>kino3434</td>\n",
       "      <td>1</td>\n",
       "      <td>8.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>visi7k1yr</td>\n",
       "      <td>1</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>megzy123</td>\n",
       "      <td>1</td>\n",
       "      <td>8.0</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>lamborghin1</td>\n",
       "      <td>1</td>\n",
       "      <td>11.0</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      password  strength  char_count  numerics  alpha  vowels  consonants\n",
       "0     kzde5577         1         8.0         4      4       1           3\n",
       "1     kino3434         1         8.0         4      4       2           2\n",
       "2    visi7k1yr         1         9.0         2      7       2           5\n",
       "3     megzy123         1         8.0         3      5       1           4\n",
       "4  lamborghin1         1        11.0         1     10       3           7"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.PairGrid at 0x2282ad08cf8>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 950.375x900 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.pairplot(data = df, hue = 'strength', vars = df.columns.difference(['password', 'strength']), \n",
    "                 diag_kind = {'kde'}, palette = 'husl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import math\n",
    "x_values = np.linspace(-5, 5, 100)\n",
    "y_values = [1 / (1 + math.e**(-x)) for x in x_values]\n",
    "plt.plot(x_values, y_values)\n",
    "plt.axhline(.5)\n",
    "plt.axvline(0)\n",
    "sns.despine()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1 / (1 + math.e**(-95))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Below is code that implements everything we discussed. It is vectorized, though, so things are represented as vectors and matricies. It should still be fairly clear what is going on (I hope...if not, please let me know and I can put out a version closer to the math). Also, I didn't implement an intercept (so no  𝛽0 ) feel free to add this if you wish :)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def logistic_func(theta, x):\n",
    "    return float(1) / (1 + math.e**(-x.dot(theta)))\n",
    "\n",
    "def log_gradient(theta, x, y):\n",
    "    first_calc = logistic_func(theta, x) - np.squeeze(y)\n",
    "    final_calc = first_calc.T.dot(x)\n",
    "    return final_calc\n",
    "\n",
    "def cost_func(theta, x, y):\n",
    "    log_func_v = logistic_func(theta,x)\n",
    "    y = np.squeeze(y)\n",
    "    step1 = y * np.log(log_func_v)\n",
    "    step2 = (1-y) * np.log(1 - log_func_v)\n",
    "    final = -step1 - step2\n",
    "    return np.mean(final)\n",
    "\n",
    "def grad_desc(theta_values, X, y, lr=.001, converge_change=.001):\n",
    "    #normalize\n",
    "    X = (X - np.mean(X, axis=0)) / np.std(X, axis=0)\n",
    "    #setup cost iter\n",
    "    cost_iter = []\n",
    "    cost = cost_func(theta_values, X, y)\n",
    "    cost_iter.append([0, cost])\n",
    "    change_cost = 1\n",
    "    i = 1\n",
    "    while(change_cost > converge_change):\n",
    "        old_cost = cost\n",
    "        theta_values = theta_values - (lr * log_gradient(theta_values, X, y))\n",
    "        cost = cost_func(theta_values, X, y)\n",
    "        cost_iter.append([i, cost])\n",
    "        change_cost = old_cost - cost\n",
    "        i+=1\n",
    "    return theta_values, np.array(cost_iter)\n",
    "\n",
    "def pred_values(theta, X, hard=True):\n",
    "    #normalize\n",
    "    X = (X - np.mean(X, axis=0)) / np.std(X, axis=0)\n",
    "    pred_prob = logistic_func(theta, X)\n",
    "    pred_value = np.where(pred_prob >= .5, 1, 0)\n",
    "    if hard:\n",
    "        return pred_value\n",
    "    return pred_prob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.dropna(inplace=True)\n",
    "X = df.iloc[:,2:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "dummy = df.strength\n",
    "d = dummy.replace(1,0)\n",
    "y = d.replace(2,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    586502\n",
       "1     83137\n",
       "Name: strength, dtype: int64"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Put it to the test\n",
    "So here I will use the above code for our example. I initalize our  𝛽  values to all be zero, then run gradient descent to learn the  𝛽  values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "char_count    175.146406\n",
      "numerics       -3.379669\n",
      "alpha         155.123504\n",
      "vowels        -20.672881\n",
      "consonants    170.519799\n",
      "dtype: float64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:12: RuntimeWarning: divide by zero encountered in log\n",
      "  if sys.path[0] == '':\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:13: RuntimeWarning: divide by zero encountered in log\n",
      "  del sys.path[0]\n"
     ]
    }
   ],
   "source": [
    "shape = X.shape[1]\n",
    "betas = np.zeros(shape)\n",
    "fitted_values, cost_iter = grad_desc(betas, X, y)\n",
    "print(fitted_values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's make some predictions (Note: since we are returning a probability, if the probability is greater than or equal to 50% then I assign the value to strong - or a value of 1):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 1, 0, 0])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predicted_y = pred_values(fitted_values, X)\n",
    "predicted_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "510693"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#And let's see how accurate we are:\n",
    "np.sum(y == predicted_y)\n",
    "#it gives 76% accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "669607"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Lastly, another nice check is to see how well a packaged version of the algorithm does:\n",
    "from sklearn import linear_model\n",
    "logreg = linear_model.LogisticRegression()\n",
    "logreg.fit(X, y)\n",
    "sum(y == logreg.predict(X))\n",
    "#99.99% accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train acc: 0.9999398205314516\n",
      "Test acc: 0.9999366461369982\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)\n",
    "clf = linear_model.LogisticRegression(penalty='l2')\n",
    "clf.fit(X_train, y_train)\n",
    "train_predictions = clf.predict(X_train)\n",
    "test_predictions = clf.predict(X_test)\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "print(\"Train acc: {}\".format(accuracy_score(y_train, train_predictions)))\n",
    "print(\"Test acc: {}\".format(accuracy_score(y_test, test_predictions)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.875848\n",
       "1    0.124152\n",
       "Name: strength, dtype: float64"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(y).value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[193478,     14],\n",
       "       [     0,  27489]], dtype=int64)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "confusion_matrix(y_test, test_predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train:\n",
      "(array([1.        , 0.99951504]), array([0.9999313, 1.       ]), array([0.99996565, 0.99975746]), array([393010,  55648], dtype=int64))\n",
      "Test:\n",
      "(array([1.        , 0.99949096]), array([0.99992765, 1.        ]), array([0.99996382, 0.99974542]), array([193492,  27489], dtype=int64))\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "print(\"Train:\")\n",
    "print(precision_recall_fscore_support(y_train, train_predictions))\n",
    "print(\"Test:\")\n",
    "print(precision_recall_fscore_support(y_test, test_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_prediction_probabilities = clf.predict_proba(X_train)\n",
    "test_prediction_probabilities = clf.predict_proba(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x22854aee978>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_probs_for_actual_1 = [train_prediction_probabilities[i][1] for i, truth in enumerate(y_train)\n",
    "                         if truth == 1]\n",
    "train_probs_for_actual_0 = [train_prediction_probabilities[i][1] for i, truth in enumerate(y_train)\n",
    "                         if truth == 0]\n",
    "plt.hist(train_probs_for_actual_1, label='1')\n",
    "plt.hist(train_probs_for_actual_0, label='0')\n",
    "plt.legend(loc='upper left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x22854b5f7f0>"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_curve, average_precision_score\n",
    "\n",
    "precisions, recalls, f1s = [], [], []\n",
    "cutoffs = np.arange(0.05, 1, 0.05)\n",
    "for cutoff in cutoffs:\n",
    "    binary_test_predictions = [x[1] >= cutoff for x in test_prediction_probabilities]\n",
    "    p, r, f1, s = precision_recall_fscore_support(y_test, binary_test_predictions)\n",
    "    precisions.append(p[1])\n",
    "    recalls.append(r[1])\n",
    "    f1s.append(f1[1])\n",
    "    \n",
    "plt.plot(cutoffs, precisions, label='precision')\n",
    "plt.plot(cutoffs, recalls, label='recall')\n",
    "plt.plot(cutoffs, f1s, label='f1')\n",
    "plt.xlabel(\"Positive Probability >=\")\n",
    "plt.title(\"Scores for Positive Class\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve, auc\n",
    "\n",
    "fpr, tpr, threshold = roc_curve(y_test, test_prediction_probabilities[:,1])\n",
    "roc_auc = auc(fpr, tpr)\n",
    "\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(fpr, tpr, 'b', label = 'AUC = %0.4f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],'r--')\n",
    "plt.xlim([0, 1])\n",
    "plt.ylim([0, 1])\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "coefficients = pd.DataFrame({\"Feature\":X.columns,\"Coefficients\":np.transpose(clf.coef_[0])})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "147.52534624995593"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(4.994)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "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>Feature</th>\n",
       "      <th>Coefficients</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>char_count</td>\n",
       "      <td>4.994563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>numerics</td>\n",
       "      <td>-0.653701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>alpha</td>\n",
       "      <td>-0.434567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>vowels</td>\n",
       "      <td>-0.450078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>consonants</td>\n",
       "      <td>0.015511</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Feature  Coefficients\n",
       "0  char_count      4.994563\n",
       "1    numerics     -0.653701\n",
       "2       alpha     -0.434567\n",
       "3      vowels     -0.450078\n",
       "4  consonants      0.015511"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coefficients"
   ]
  },
  {
   "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
}