loan_prediction_using_other model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading Packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np                     # For mathematical calculations \n",
    "import seaborn as sns                  # For data visualization \n",
    "import matplotlib.pyplot as plt        # For plotting graphs \n",
    "%matplotlib inline \n",
    "import warnings                        # To ignore any warnings warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train =pd.read_csv('train_final.csv')\n",
    "test = pd.read_csv('test_final.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Credit_History', 'Loan_Status', 'LoanAmount_log', 'Gender_Female',\n",
       "       'Gender_Male', 'Married_No', 'Married_Yes', 'Dependents_3',\n",
       "       'Dependents_0', 'Dependents_1', 'Dependents_2', 'Education_Graduate',\n",
       "       'Education_Not Graduate', 'Self_Employed_No', 'Self_Employed_Yes',\n",
       "       'Property_Area_Rural', 'Property_Area_Semiurban', 'Property_Area_Urban',\n",
       "       'Total_Income_log', 'EMI', 'Balance Income'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.columns"
   ]
  },
  {
   "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>Credit_History</th>\n",
       "      <th>Loan_Status</th>\n",
       "      <th>LoanAmount_log</th>\n",
       "      <th>Gender_Female</th>\n",
       "      <th>Gender_Male</th>\n",
       "      <th>Married_No</th>\n",
       "      <th>Married_Yes</th>\n",
       "      <th>Dependents_3</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>...</th>\n",
       "      <th>Education_Graduate</th>\n",
       "      <th>Education_Not Graduate</th>\n",
       "      <th>Self_Employed_No</th>\n",
       "      <th>Self_Employed_Yes</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "      <th>Total_Income_log</th>\n",
       "      <th>EMI</th>\n",
       "      <th>Balance Income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.852030</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8.674026</td>\n",
       "      <td>0.355556</td>\n",
       "      <td>5493.444444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.852030</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.714568</td>\n",
       "      <td>0.355556</td>\n",
       "      <td>5735.444444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.189655</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8.006368</td>\n",
       "      <td>0.183333</td>\n",
       "      <td>2816.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.787492</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8.505323</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>4607.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4.948760</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8.699515</td>\n",
       "      <td>0.391667</td>\n",
       "      <td>5608.333333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Credit_History  Loan_Status  LoanAmount_log  Gender_Female  Gender_Male  \\\n",
       "0             1.0            1        4.852030              0            1   \n",
       "1             1.0            0        4.852030              0            1   \n",
       "2             1.0            1        4.189655              0            1   \n",
       "3             1.0            1        4.787492              0            1   \n",
       "4             1.0            1        4.948760              0            1   \n",
       "\n",
       "   Married_No  Married_Yes  Dependents_3  Dependents_0  Dependents_1  ...  \\\n",
       "0           1            0             0             1             0  ...   \n",
       "1           0            1             0             0             1  ...   \n",
       "2           0            1             0             1             0  ...   \n",
       "3           0            1             0             1             0  ...   \n",
       "4           1            0             0             1             0  ...   \n",
       "\n",
       "   Education_Graduate  Education_Not Graduate  Self_Employed_No  \\\n",
       "0                   1                       0                 1   \n",
       "1                   1                       0                 1   \n",
       "2                   1                       0                 0   \n",
       "3                   0                       1                 1   \n",
       "4                   1                       0                 1   \n",
       "\n",
       "   Self_Employed_Yes  Property_Area_Rural  Property_Area_Semiurban  \\\n",
       "0                  0                    0                        0   \n",
       "1                  0                    1                        0   \n",
       "2                  1                    0                        0   \n",
       "3                  0                    0                        0   \n",
       "4                  0                    0                        0   \n",
       "\n",
       "   Property_Area_Urban  Total_Income_log       EMI  Balance Income  \n",
       "0                    1          8.674026  0.355556     5493.444444  \n",
       "1                    0          8.714568  0.355556     5735.444444  \n",
       "2                    1          8.006368  0.183333     2816.666667  \n",
       "3                    1          8.505323  0.333333     4607.666667  \n",
       "4                    1          8.699515  0.391667     5608.333333  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Credit_History             0\n",
       "LoanAmount_log             0\n",
       "Gender_Female              0\n",
       "Gender_Male                0\n",
       "Married_No                 0\n",
       "Married_Yes                0\n",
       "Dependents_3               0\n",
       "Dependents_0               0\n",
       "Dependents_1               0\n",
       "Dependents_2               0\n",
       "Education_Graduate         0\n",
       "Education_Not Graduate     0\n",
       "Self_Employed_No           0\n",
       "Self_Employed_Yes          0\n",
       "Property_Area_Rural        0\n",
       "Property_Area_Semiurban    0\n",
       "Property_Area_Urban        0\n",
       "Total_Income_log           0\n",
       "EMI                        0\n",
       "Balance Income             0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "train['EMI'].fillna(train['EMI'].median(), inplace=True)\n",
    "train['Balance Income'].fillna(train['Balance Income'].median(), inplace=True)\n",
    "\n",
    "test['EMI'].fillna(test['EMI'].median(), inplace=True)\n",
    "test['Balance Income'].fillna(test['Balance Income'].median(), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = train.drop('Loan_Status',1) \n",
    "y = train.Loan_Status"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Credit_History</th>\n",
       "      <th>LoanAmount_log</th>\n",
       "      <th>Gender_Female</th>\n",
       "      <th>Gender_Male</th>\n",
       "      <th>Married_No</th>\n",
       "      <th>Married_Yes</th>\n",
       "      <th>Dependents_3</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Education_Graduate</th>\n",
       "      <th>Education_Not Graduate</th>\n",
       "      <th>Self_Employed_No</th>\n",
       "      <th>Self_Employed_Yes</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "      <th>Total_Income_log</th>\n",
       "      <th>EMI</th>\n",
       "      <th>Balance Income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.85203</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8.674026</td>\n",
       "      <td>0.355556</td>\n",
       "      <td>5493.444444</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Credit_History  LoanAmount_log  Gender_Female  Gender_Male  Married_No  \\\n",
       "0             1.0         4.85203              0            1           1   \n",
       "\n",
       "   Married_Yes  Dependents_3  Dependents_0  Dependents_1  Dependents_2  \\\n",
       "0            0             0             1             0             0   \n",
       "\n",
       "   Education_Graduate  Education_Not Graduate  Self_Employed_No  \\\n",
       "0                   1                       0                 1   \n",
       "\n",
       "   Self_Employed_Yes  Property_Area_Rural  Property_Area_Semiurban  \\\n",
       "0                  0                    0                        0   \n",
       "\n",
       "   Property_Area_Urban  Total_Income_log       EMI  Balance Income  \n",
       "0                    1          8.674026  0.355556     5493.444444  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use the train_test_split function from sklearn to divide our train dataset. So, first let us import train_test_split."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_cv, y_train, y_cv = train_test_split(X,y, test_size =0.3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will train the model on training dataset and make predictions for the test dataset. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Decision Tree\n",
    "Decision tree is a type of supervised learning algorithm(having a pre-defined target variable) that is mostly used in classification problems. In this technique, we split the population or sample into two or more homogeneous sets(or sub-populations) based on most significant splitter / differentiator in input variables.\n",
    "\n",
    "Decision trees use multiple algorithms to decide to split a node in two or more sub-nodes. The creation of sub-nodes increases the homogeneity of resultant sub-nodes. In other words, we can say that purity of the node increases with respect to the target variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(random_state=1)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "model = DecisionTreeClassifier(random_state=1)  \n",
    "model.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let’s predict the Loan_Status for validation set and calculate its accuracy.\n",
    "\n",
    "pred_cv = model.predict(x_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7351351351351352"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Let us calculate how accurate our predictions are by calculating the accuracy.\n",
    "\n",
    "accuracy_score(y_cv,pred_cv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To check how robust our model is to unseen data, we can use Validation. It is a technique which involves reserving a particular sample of a dataset on which you do not train the model. Later, you test your model on this sample before finalizing it. Some of the common methods for validation are listed below:\n",
    "\n",
    "1. The validation set approach\n",
    "2. k-fold cross validation\n",
    "3. Leave one out cross validation (LOOCV)\n",
    "4. Stratified k-fold cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "1 of kfold 5\n",
      "accuracy_score 0.7235772357723578\n",
      "\n",
      "2 of kfold 5\n",
      "accuracy_score 0.6910569105691057\n",
      "\n",
      "3 of kfold 5\n",
      "accuracy_score 0.7723577235772358\n",
      "\n",
      "4 of kfold 5\n",
      "accuracy_score 0.7317073170731707\n",
      "\n",
      "5 of kfold 5\n",
      "accuracy_score 0.680327868852459\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import StratifiedKFold\n",
    "\n",
    "i=1 \n",
    "kf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \n",
    "for train_index,test_index in kf.split(X,y):     \n",
    "    print('\\n{} of kfold {}'.format(i,kf.n_splits))  \n",
    "    xtr,xvl = X.loc[train_index],X.loc[test_index]  \n",
    "    ytr,yvl = y[train_index],y[test_index]  \n",
    "    model = DecisionTreeClassifier(random_state=1)   \n",
    "    model.fit(xtr, ytr) \n",
    "    pred_test = model.predict(xvl) \n",
    "    score = accuracy_score(yvl,pred_test)  \n",
    "    print('accuracy_score',score)  \n",
    "    i+=1 \n",
    "    pred_test = model.predict(test)\n",
    "    pred=model.predict_proba(xvl)[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import metrics \n",
    "fpr, tpr, _ = metrics.roc_curve(yvl,  pred) \n",
    "auc = metrics.roc_auc_score(yvl, pred) \n",
    "plt.figure(figsize=(12,8)) \n",
    "plt.plot(fpr,tpr,label=\"validation, auc=\"+str(auc))\n",
    "plt.xlabel('False Positive Rate') \n",
    "plt.ylabel('True Positive Rate') \n",
    "plt.legend(loc=4)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forest\n",
    "RandomForest is a tree based bootstrapping algorithm wherein a certain no. of weak learners (decision trees) are combined to make a powerful prediction model.\n",
    "For every individual learner, a random sample of rows and a few randomly chosen variables are used to build a decision tree model.\n",
    "Final prediction can be a function of all the predictions made by the individual learners.\n",
    "In case of regression problem, the final prediction can be mean of all the predictions.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "1 of kfold 5\n",
      "accuracy_score 0.8292682926829268\n",
      "\n",
      "2 of kfold 5\n",
      "accuracy_score 0.8211382113821138\n",
      "\n",
      "3 of kfold 5\n",
      "accuracy_score 0.7967479674796748\n",
      "\n",
      "4 of kfold 5\n",
      "accuracy_score 0.8130081300813008\n",
      "\n",
      "5 of kfold 5\n",
      "accuracy_score 0.7786885245901639\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "i=1 \n",
    "kf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \n",
    "for train_index,test_index in kf.split(X,y):     \n",
    "    print('\\n{} of kfold {}'.format(i,kf.n_splits))  \n",
    "    xtr,xvl = X.loc[train_index],X.loc[test_index]  \n",
    "    ytr,yvl = y[train_index],y[test_index]  \n",
    "    model = RandomForestClassifier(random_state=1, max_depth=10)   \n",
    "    model.fit(xtr, ytr) \n",
    "    pred_test = model.predict(xvl) \n",
    "    score = accuracy_score(yvl,pred_test)  \n",
    "    print('accuracy_score',score)  \n",
    "    i+=1 \n",
    "    pred_test = model.predict(test)\n",
    "    pred=model.predict_proba(xvl)[:,1]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import metrics \n",
    "fpr, tpr, _ = metrics.roc_curve(yvl,  pred) \n",
    "auc = metrics.roc_auc_score(yvl, pred) \n",
    "plt.figure(figsize=(12,8)) \n",
    "plt.plot(fpr,tpr,label=\"validation, auc=\"+str(auc))\n",
    "plt.xlabel('False Positive Rate') \n",
    "plt.ylabel('True Positive Rate') \n",
    "plt.legend(loc=4)\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}