TCC_II-Code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "quwDxZbI-URI"
   },
   "source": [
    "# TCC2\n",
    "\n",
    "## Engenharia de Sistemas - UFMG\n",
    "---\n",
    "Autor: Thullio Debortoli Moreira Zanetti\n",
    "\n",
    "Orientador: Frederico Gadelha Guimarães"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "3oJ9eNTJtZxT"
   },
   "source": [
    "## 1) Limpar  e transformar base de dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 191
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 32520,
     "status": "ok",
     "timestamp": 1540944328599,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "https://lh6.googleusercontent.com/-c5y_XqvyJ1I/AAAAAAAAAAI/AAAAAAAAAsQ/KG1ekcjD0uk/s64/photo.jpg",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "y1cisVsC1uei",
    "outputId": "ef60fd28-c04f-4aae-936a-0fa914636c2f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
      "\n",
      "Enter your authorization code:\n",
      "··········\n",
      "Mounted at /content/drive\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "J73AltXr3sa1"
   },
   "outputs": [],
   "source": [
    "import sqlite3\n",
    "import pandas as pd \n",
    "import numpy as np\n",
    "#from google.colab import files\n",
    "import matplotlib.pylab as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "HbRyk8Ci3-d6"
   },
   "outputs": [],
   "source": [
    "#df = pd.read_csv(files.download(\"drive/UFMG/TCC2/Algoritmo/database/loan.csv\"), sep=',', usecols = [\"loan_status\", \"loan_amnt\", \"term\", \"int_rate\"]) \n",
    "#['Fully Paid', 'Charged Off', 'Current', 'Default',\n",
    "#       'Late (31-120 days)', 'In Grace Period', 'Late (16-30 days)',\n",
    "#       'Does not meet the credit policy. Status:Fully Paid',\n",
    "#       'Does not meet the credit policy. Status:Charged Off', 'Issued']\n",
    "conn = sqlite3.connect('/content/drive/My Drive/UFMG/TCC2/Algoritmo/database/database.sqlite')\n",
    "to_parse = ['issue_d' , 'earliest_cr_line', 'last_pymnt_d', 'next_pymnt_d', 'last_credit_pull_d']\n",
    "df_db = pd.read_sql_query(\"select * from loan where loan_status = 'Fully Paid' OR loan_status = 'Charged Off' OR loan_status = 'Default'\", con=conn, parse_dates = to_parse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "e497eW2gjW4v"
   },
   "source": [
    "Pimeiramente são excluidas as colunas e as linhas que contém mais de 50% dos seus valores nulos, pois não possuem informação suficiente para o problema"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "2m2EHSFv-3y1"
   },
   "outputs": [],
   "source": [
    "df = df_db\n",
    "df.dropna(axis=1, thresh=int(df.shape[0]*0.5), inplace=True)\n",
    "df.dropna(axis=0, thresh=int(df.shape[1]*0.5), inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "0Dt2jbL1kSkX"
   },
   "source": [
    "Das colunas restantes, como para predição são utilizados apenas dados coletados anteriormente à concessão do empréstimo, algumas colunas não possuem sentido para o problema. Outras são campos aberto e também não agragam valor.\n",
    "\n",
    "\n",
    "*   Colunas excluídas: index, id, member_id, issue_d, initial_list_status, total_pymnt, total_pymnt_inv, total_rec_prncp, total_rec_int, total_rec_late_fee, collection_recovery_fee, last_pymnt_d, last_pymnt_amnt, last_credit_pull_d, collections_12_mths_ex_med, pymnt_plan, url, zip_code, addr_state, grade, emp_title, purpose, policy_code, title\n",
    "\n",
    "*   Colunas mantidas: loan_amnt, funded_amnt, funded_amnt_inv, term, int_rate, installment, sub_grade, emp_length, home_ownership, annual_inc, verification_status, loan_status, dti, delinq_2yrs, earliest_cr_line, inq_last_6mths, open_acc, pub_rec, revol_bal, revol_util, total_acc, recoveries,  application_type, acc_now_delinq, tot_coll_amt, tot_cur_bal, total_rev_hi_lim\n",
    "\n",
    "Depois verifica-se quais colunas são do tipo texto para que sejam transformadas.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "rtYqAb5OC7Bs"
   },
   "outputs": [],
   "source": [
    "columns_to_keep =  ['loan_amnt',\n",
    "                    'funded_amnt',\n",
    "                    'funded_amnt_inv',\n",
    "                    'term',\n",
    "                    'int_rate',\n",
    "                    'installment',\n",
    "                    'sub_grade',\n",
    "                    'emp_length',\n",
    "                    'home_ownership',\n",
    "                    'annual_inc',\n",
    "                    'verification_status',\n",
    "                    'loan_status',\n",
    "                    'dti',\n",
    "                    'delinq_2yrs',\n",
    "                    'earliest_cr_line',\n",
    "                    'inq_last_6mths',\n",
    "                    'open_acc',\n",
    "                    'pub_rec',\n",
    "                    'revol_bal',\n",
    "                    'revol_util',\n",
    "                    'total_acc',\n",
    "                    'recoveries',\n",
    "                    'application_type',\n",
    "                    'acc_now_delinq',\n",
    "                    'tot_coll_amt',\n",
    "                    'tot_cur_bal',\n",
    "                    'total_rev_hi_lim']\n",
    "df = df[columns_to_keep]\n",
    "#df.dtypes  #get column types"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SZSo6zrLkEkP"
   },
   "source": [
    "**Segue relação de colunas e transformações a serem realizadas:**\n",
    "\n",
    "Coluna | Ex valor atual | Ex novo valor | Novo nome\n",
    ":---: | :---: | :---: | :---:\n",
    "term |  36 m | 36 | -\n",
    "int_rate | 10% | 0.1 | -\n",
    "sub_grade | [A1 - G5] | [35 - 1] | grade\n",
    "emp_length | 1 year | 1 | -\n",
    "home_ownership | own/other | 1/0 |  own_home              \n",
    "verification_status | Verified/Not Verified | 1/0 | verified\n",
    "loan_status | Fully Paid/Charged off | 1/0 | fully_paid\n",
    "earliest_cr_line | DateTime | Month/Year |  -\n",
    "revol_util | 10% | 0.1 | -\n",
    "application_type | INDIVIDUAL/JOINT | 1/0 | individual"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1193
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 6563,
     "status": "ok",
     "timestamp": 1540944395707,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "https://lh6.googleusercontent.com/-c5y_XqvyJ1I/AAAAAAAAAAI/AAAAAAAAAsQ/KG1ekcjD0uk/s64/photo.jpg",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "LvutIvcykGsO",
    "outputId": "a1cadced-bd17-40ca-bfad-263f257f8dab"
   },
   "outputs": [],
   "source": [
    "#Define range transformation\n",
    "grades = {'A1' : 35, 'A2' : 34, 'A3' : 33, 'A4' : 32, 'A5' : 31, 'B1' : 30, 'B2' : 29, 'B3' : 28, 'B4' : 27, 'B5' : 26, 'C1' : 25, 'C2' : 24, 'C3' : 23, 'C4' : 22, 'C5' : 21, 'D1' : 20, 'D2' : 19, 'D3' : 18, 'D4' : 17, 'D5' : 16, 'E1' : 15, 'E2' : 14, 'E3' : 13, 'E4' : 12, 'E5' : 11, 'F1' : 10, 'F2' : 9, 'F3' : 8, 'F4' : 7, 'F5' : 6, 'G1' : 5, 'G2' : 4, 'G3' : 3, 'G4' : 2, 'G5' : 1}\n",
    "\n",
    "#applying transformations\n",
    "df['term'] = df['term'].str.split(' ').str[1]\n",
    "df['int_rate'] = (df['int_rate'].str.split('%').str[0]).astype(float)/100.\n",
    "df[\"sub_grade\"] = df[\"sub_grade\"].replace(grades)\n",
    "# extract numbers from emp_length and fill missing values with the median\n",
    "df['emp_length'] = df['emp_length'].str.extract('(\\d+)').astype(float)\n",
    "df['emp_length'] = df['emp_length'].fillna(df.emp_length.median())\n",
    "df['home_ownership'] = df['home_ownership'].apply(lambda x : 1 if x == 'OWN' else 0)\n",
    "df['verification_status'] = df['verification_status'].apply(lambda x : 0 if x == 'Not Verified' else 1)\n",
    "df['loan_status'] = df['loan_status'].apply(lambda x : 1 if x == 'Fully Paid' else 0)\n",
    "df['earliest_cr_line'] = df['earliest_cr_line'].dt.to_period('M')\n",
    "df['revol_util'] = (df['revol_util'].str.split('%').str[0]).astype(float)/100.\n",
    "df['application_type'] = df['application_type'].apply(lambda x : 1 if x == 'INDIVIDUAL' else 0)\n",
    "\n",
    "df = df.rename(index=str, columns={\"loan_status\": \"fully_paid\", \n",
    "                              \"sub_grade\": \"grade\", \n",
    "                              \"home_ownership\" : \"own_home\", \n",
    "                              \"verification_status\" : \"verified\", \n",
    "                              \"application_type\" : \"individual\"})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "w1hVoI6k6ski"
   },
   "source": [
    "A partir das características restantes, foram criadas novas caracteristicas em busca de uma informação melhor. As colunas funded_amnt e funded_amnt_inv foram substituídas pelo percentual dos valores em relação ao montante do empréstimo."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "OXtvFFhp6rvT"
   },
   "outputs": [],
   "source": [
    "df['funded_amnt'] = df['funded_amnt']/df['loan_amnt']\n",
    "df['funded_amnt_inv'] = df['funded_amnt_inv']/df['loan_amnt']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "7i_-W3U9y2dF"
   },
   "source": [
    "Após a transformação dos dados, os mesmos são filtrados de acordo com a correlação dos mesmos com a saída. São retirados as variáveis com valor absoluto de correlação menor que 0,01 com a saída. Após este filtro, retaram 15 parâmetros de entrada finais."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZBm3Dmzzvy5H"
   },
   "outputs": [],
   "source": [
    "df = df.fillna(0)\n",
    "df = df[df.corr()['fully_paid'][np.absolute(df.corr()['fully_paid']) > 0.01].index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 102
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1526,
     "status": "ok",
     "timestamp": 1540944518181,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "https://lh6.googleusercontent.com/-c5y_XqvyJ1I/AAAAAAAAAAI/AAAAAAAAAsQ/KG1ekcjD0uk/s64/photo.jpg",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "rbMQJrNOczWH",
    "outputId": "de3cddf3-1bae-4cc1-dce0-25b20284a427"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['loan_amnt', 'int_rate', 'installment', 'grade', 'annual_inc',\n",
       "       'verified', 'fully_paid', 'dti', 'delinq_2yrs', 'inq_last_6mths',\n",
       "       'open_acc', 'revol_util', 'total_acc', 'recoveries', 'tot_cur_bal',\n",
       "       'total_rev_hi_lim'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 9,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ftvOLuWnwVWV"
   },
   "source": [
    "# 2) Modelos de IA\n",
    "\n",
    "10% do dataset sera separado para testes e comparação dos resultados. Todos os modelos serão treinados utilizando o método K-fold dividindo-se o dataset em 9 partes iguais.\n",
    "\n",
    "## 2.1) Decision Tree\n",
    "\n",
    "A primeira técnica de inteligência artificial a ser aplicada é a de uma árvore de decisão."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "fkS0zXTd0D23"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn import tree\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "import pickle\n",
    "from sklearn.externals import joblib\n",
    "from sklearn import metrics\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "VfkHfwPCwdZM"
   },
   "outputs": [],
   "source": [
    "#Separar Entrada e Saída\n",
    "y = df['fully_paid']\n",
    "X = df.drop(['fully_paid'], axis=1)\n",
    "errorI = []\n",
    "errorII = []\n",
    "accuracy = []\n",
    "roc_auc_score = []\n",
    "out_test = []\n",
    "out_prob = []\n",
    "\n",
    "for t in range(0, 30):\n",
    "  # Define entrada e saída para treinamento e teste (10% da amostra)\n",
    "  '''X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1)\n",
    "\n",
    "  #Treinar modelo utilizando GridSearchCV e variando profundidade da arvore de 3 a 20, com 9 folds\n",
    "  #Algoritmo utilizado para gerar o classificador:\n",
    "  parameters = {'max_depth':range(3,20)}\n",
    "  decTreeClf = GridSearchCV(tree.DecisionTreeClassifier(), parameters, n_jobs=4, cv = 9)\n",
    "  decTreeClf.fit(X=X_train, y=y_train)\n",
    "  \n",
    "  decTreeClf.X_train = X_train\n",
    "  decTreeClf.X_test = X_test\n",
    "  decTreeClf.y_train = y_train\n",
    "  decTreeClf.y_test = y_test\n",
    "\n",
    "  with open('/content/drive/My Drive/UFMG/TCC2/Algoritmo/classifiers/Decision Tree without augmentation/decisionTreeCLF{}.pkl'.format(t), 'wb') as f:\n",
    "      pickle.dump(decTreeClf, f) # save the object to a file'''\n",
    "\n",
    "  with open('/content/drive/My Drive/UFMG/TCC2/Algoritmo/classifiers/Decision Tree without augmentation/decisionTreeCLF{}.pkl'.format(t), 'rb') as f:\n",
    "      decTreeClf = pickle.load(f)\n",
    "      \n",
    "  X_train = decTreeClf.X_train\n",
    "  X_test = decTreeClf.X_test\n",
    "  y_train = decTreeClf.y_train\n",
    "  y_test = decTreeClf.y_test\n",
    "  \n",
    "  #Testar modelo obtido\n",
    "  y_predict_tree = decTreeClf.predict(X_test)\n",
    "  y_difference_tree = y_test - y_predict_tree\n",
    "  y_pred_proba = decTreeClf.predict_proba(X_test)[::,1]\n",
    "  out_test.extend(y_test)\n",
    "  out_prob.extend(y_pred_proba)\n",
    "  errorI.append(float(y_difference_tree[y_difference_tree == -1].shape[0])/float(y_difference_tree.shape[0]))\n",
    "  errorII.append(float(y_difference_tree[y_difference_tree == 1].shape[0])/float(y_difference_tree.shape[0]))\n",
    "  accuracy.append(float(y_difference_tree[y_difference_tree == 0].shape[0])/float(y_difference_tree.shape[0]))\n",
    "  roc_auc_score.append(metrics.roc_auc_score(y_test, y_pred_proba))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 402
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1169,
     "status": "ok",
     "timestamp": 1537579279687,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "9mUuWjgTAWjq",
    "outputId": "cae230ad-a9d2-4379-ef89-037a6909d652"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Type I error (FP): 0.08780\n",
      "Mean Type II error (FN): 0.00000\n",
      "Mean Accuracy: 0.91220\n",
      "Mean ROC_AUC_Score: 0.83788\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEbCAYAAADajfNFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f231093c240>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print ('Mean Type I error (FP): {0:.5f}'.format(np.mean(errorI)))\n",
    "print ('Mean Type II error (FN): {0:.5f}'.format(np.mean(errorII)))\n",
    "print ('Mean Accuracy: {0:.5f}'.format(np.mean(accuracy)))\n",
    "print ('Mean ROC_AUC_Score: {0:.5f}\\n\\n'.format(np.mean(roc_auc_score)))\n",
    "#print 'Mean Depth: {0:.5f}\\n\\n'.format(np.mean(depth))\n",
    "\n",
    "idx_max = np.argmax(accuracy)\n",
    "\n",
    "\n",
    "y_test = out_test[int(idx_max*(len(out_test)/30)):int((idx_max*(len(out_test)/30))+(len(out_test)/30))]\n",
    "y_pred_proba = out_prob[int(idx_max*(len(out_test)/30)):int((idx_max*(len(out_test)/30))+(len(out_test)/30))]\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test,  y_pred_proba)\n",
    "plt.plot(fpr,tpr, linewidth=2)\n",
    "plt.xlabel(\"False Positive Rate\", fontsize=14)\n",
    "plt.ylabel(\"True Positive Rate\", fontsize=14)\n",
    "plt.title(\"ROC Curve\", fontsize=14)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7n-e7QEYBref"
   },
   "outputs": [],
   "source": [
    "# Codigo para conseguir printar datas\n",
    "%%bash\n",
    "/usr/bin/yes | apt-get install graphviz\n",
    "pip install --upgrade graphviz\n",
    "/usr/bin/yes | pip uninstall pyparsing\n",
    "pip install -Iv https://pypi.python.org/packages/source/p/pyparsing/pyparsing-1.5.7.tar.gz\n",
    "pip install --upgrade pydot\n",
    "pip install pydotplus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 514
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1809,
     "status": "ok",
     "timestamp": 1537579303814,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "CfQb6uDdlc2n",
    "outputId": "0a47a1d2-f63c-4b42-f793-7089cf8a3b11"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAgAAAHxCAIAAABvXhk7AAAABmJLR0QA/wD/AP+gvaeTAAAgAElE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\nMYCBgYGBgYGBgYGBAQD8D00hC3s2VR0BAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 27,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open('/content/drive/My Drive/UFMG/TCC2/Algoritmo/classifiers/Decision Tree without augmentation/decisionTreeCLF{}.pkl'.format(idx_max), 'rb') as f:\n",
    "  decTreeClf = pickle.load(f)\n",
    "\n",
    "tree_model = decTreeClf.best_estimator_\n",
    "\n",
    "# Draw Decision Tree\n",
    "from sklearn.externals.six import StringIO  \n",
    "from IPython.display import Image  \n",
    "from sklearn.tree import export_graphviz\n",
    "import pydotplus\n",
    "dot_data = StringIO()\n",
    "export_graphviz(tree_model, out_file=dot_data, feature_names=X.columns, \n",
    "                filled=True, rounded=True,\n",
    "                special_characters=True)\n",
    "graph = pydotplus.graph_from_dot_data(dot_data.getvalue())  \n",
    "Image(graph.create_png())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "mlHdJxpgfCG1"
   },
   "source": [
    "### 2.1.1) Desbalanceamento de dados\n",
    "Os resultados do primeiro modelo de Decision Tree não foram satisfatórios, então foi buscado um motivo para tal resultado. A justificativa para ter um modelo que leva apenas uma variável em consideração é o desbalancemaneto de dados, no qual a parcela de dados de clientes que pagaram integralmente o empréstimo é de  81.72% dos dados totais. Para isto, será empregado uma técnica de oversampling no treinamento dos modelos, para que equalize as classes.\n",
    "\n",
    "Para isso,também será criado uma função para separar os dados de teste de forma balanceada, usando 5.48% dos dados totais, dos quais metade serão da classe *inadimplentes*, o que representa 15% da classe. Este número foi escolhido para que não ocorra perda de características da classe para o treinamento. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "R35r0OoGlQEJ"
   },
   "outputs": [],
   "source": [
    "from sklearn.utils import shuffle\n",
    "\n",
    "def balanced_subsample(x, y, subsample_minority_class_size=0.15):\n",
    "\n",
    "    class_xs = []\n",
    "    min_elems = None\n",
    "\n",
    "    for yi in np.unique(y):\n",
    "        elems = x[(y == yi)]\n",
    "        class_xs.append((yi, elems))\n",
    "        if min_elems == None or elems.shape[0] < min_elems:\n",
    "            min_elems = elems.shape[0]\n",
    "\n",
    "    use_elems = min_elems\n",
    "    if subsample_minority_class_size < 1:\n",
    "        use_elems = int(min_elems*subsample_minority_class_size)\n",
    "\n",
    "    x_test = []\n",
    "    y_test = []\n",
    "    x_train = []\n",
    "    y_train = []\n",
    "\n",
    "    for ci,this_xs in class_xs:\n",
    "        if len(this_xs) > use_elems:\n",
    "            this_xs = shuffle(this_xs)\n",
    "\n",
    "        xs_test = this_xs[:use_elems]\n",
    "        ys_test = np.empty(use_elems)\n",
    "        ys_test.fill(ci)\n",
    "        \n",
    "        xs_train = this_xs[use_elems:]\n",
    "        ys_train = np.empty(xs_train.shape[0])\n",
    "        ys_train.fill(ci)\n",
    "\n",
    "        x_test.append(xs_test)\n",
    "        y_test.append(ys_test)\n",
    "        \n",
    "        x_train.append(xs_train)\n",
    "        y_train.append(ys_train)\n",
    "\n",
    "    x_test, y_test = shuffle(np.concatenate(x_test), np.concatenate(y_test), random_state=0)\n",
    "    x_train, y_train = shuffle(np.concatenate(x_train), np.concatenate(y_train), random_state=0)\n",
    "\n",
    "    return x_train, x_test, y_train, y_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "uGefyCQ0r8MC"
   },
   "source": [
    "### 2.1.2) Decision Tree with oversampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "bRhsEOqC_l-e"
   },
   "outputs": [],
   "source": [
    "!pip uninstall imbalanced-learn\n",
    "!pip install imbalanced-learn==0.3.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "x8xK3dL8-WhR"
   },
   "outputs": [],
   "source": [
    "import imblearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_LO63nJwsgeK"
   },
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "wD7HeY29mClp"
   },
   "outputs": [],
   "source": [
    "def get_clf_with_oversampling(X, y, path, train, parameters, method, save_method):\n",
    "  errorI = []\n",
    "  errorII = []\n",
    "  accuracy = []\n",
    "  roc_auc_score = []\n",
    "  out_test = []\n",
    "  out_prob = []\n",
    "\n",
    "  for t in range(0, 30):\n",
    "    if train:\n",
    "      # Define entrada e saída para treinamento e teste\n",
    "      X_train, X_test, y_train, y_test = balanced_subsample(X, y)\n",
    "\n",
    "      #Oversampling\n",
    "      sm = SMOTE(random_state=42)\n",
    "      X_train, y_train = sm.fit_sample(X_train, y_train)\n",
    "\n",
    "      #Algoritmo utilizado para gerar o classificador:\n",
    "      clf = GridSearchCV(method, parameters, n_jobs=4, cv = 9, verbose=5)\n",
    "      clf.fit(X=X_train, y=y_train)\n",
    "\n",
    "      clf.X_train = X_train\n",
    "      clf.X_test = X_test\n",
    "      clf.y_train = y_train\n",
    "      clf.y_test = y_test\n",
    "      \n",
    "      if save_method == 'sav':\n",
    "        with open('{}{}.sav'.format(path, t), 'wb') as f:\n",
    "            joblib.dump(clf, f) # save the object to a file\n",
    "            \n",
    "      else:\n",
    "        with open('{}{}.pkl'.format(path, t), 'wb') as f:\n",
    "          clf = pickle.dump(clf, f)\n",
    "          \n",
    "    else:\n",
    "      if save_method == 'sav':\n",
    "        with open('{}{}.sav'.format(path, t), 'rb') as f:\n",
    "            clf = joblib.load(f)\n",
    "            \n",
    "      else:\n",
    "        with open('{}{}.pkl'.format(path, t), 'rb') as f:\n",
    "          clf = pickle.load(f)\n",
    "      \n",
    "\n",
    "      X_train = clf.X_train\n",
    "      X_test = clf.X_test\n",
    "      y_train = clf.y_train\n",
    "      y_test = clf.y_test\n",
    "\n",
    "    #Testar modelo obtido\n",
    "    y_predict_tree = clf.predict(X_test)\n",
    "    y_difference_tree = y_test - y_predict_tree\n",
    "    y_pred_proba = clf.predict_proba(X_test)[::,1]\n",
    "    out_test.extend(y_test)\n",
    "    out_prob.extend(y_pred_proba)\n",
    "    errorI.append(float(y_difference_tree[y_difference_tree == -1].shape[0])/float(y_difference_tree.shape[0]))\n",
    "    errorII.append(float(y_difference_tree[y_difference_tree == 1].shape[0])/float(y_difference_tree.shape[0]))\n",
    "    accuracy.append(float(y_difference_tree[y_difference_tree == 0].shape[0])/float(y_difference_tree.shape[0]))\n",
    "    roc_auc_score.append(metrics.roc_auc_score(y_test, y_pred_proba))\n",
    "    \n",
    "  print ('Mean Type I error (FP): {0:.5f}'.format(np.mean(errorI)))\n",
    "  print ('Mean Type II error (FN): {0:.5f}'.format(np.mean(errorII)))\n",
    "  print ('Mean Accuracy: {0:.5f}'.format(np.mean(accuracy)))\n",
    "  print ('Mean ROC_AUC_Score: {0:.5f}\\n\\n'.format(np.mean(roc_auc_score)))\n",
    "  #print 'Mean Depth: {0:.5f}\\n\\n'.format(np.mean(depth))\n",
    "\n",
    "  idx_max = np.argmax(accuracy)\n",
    "\n",
    "\n",
    "  y_test = out_test[int(idx_max*(len(out_test)/30)):int((idx_max*(len(out_test)/30))+(len(out_test)/30))]\n",
    "  y_pred_proba = out_prob[int(idx_max*(len(out_test)/30)):int((idx_max*(len(out_test)/30))+(len(out_test)/30))]\n",
    "\n",
    "  fpr, tpr, _ = metrics.roc_curve(y_test,  y_pred_proba)\n",
    "  plt.plot(fpr,tpr, linewidth=2)\n",
    "  plt.xlabel(\"False Positive Rate\", fontsize=14)\n",
    "  plt.ylabel(\"True Positive Rate\", fontsize=14)\n",
    "  plt.title(\"ROC Curve\", fontsize=14)\n",
    "  plt.show()\n",
    "  \n",
    "  return idx_max"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "YlfK0ABCvikM",
    "outputId": "a835edf4-3651-45bb-9668-29557baf6548"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Type I error (FP): 0.23303\n",
      "Mean Type II error (FN): 0.00987\n",
      "Mean Accuracy: 0.75710\n",
      "Mean ROC_AUC_Score: 0.83447\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAEbCAYAAAAvc3j1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xd8HNW5//HPI9ly7w3jJlfAmGIsjA2EcjGhhEAKIaa3\nYCCQckmDm5AQSG4S+KVAQqgB0+sNYIIJoXcXmWJsg0G4yja4y7JlWe35/TEjs5Z3pZG9Rbv6vl+v\nfWln5uzMM5I9z545Z84xd0dERCSevEwHICIiLZeShIiIJKQkISIiCSlJiIhIQkoSIiKSkJKEiIgk\npCQhIiIJKUlITjCzqWbm4avGzJaZ2S1m1iNO2UPMbJqZrTezbWb2kZn9yszaxyl7oJk9YmafmVml\nmZWEx9qviXj6mdmNZvZpeIwVZvasmZ2YzPMWSTUlCcklLwD9gULgO8BXgb/HFjCzk4HXgXXAJGAU\n8GtgCvAfMyuIKXsSMBPoDJwN7ANMBlYBv08UhJkVAu8AxwFXAfuHx3oGuHV3TjA2PpG0cHe99Mr6\nFzAV+FeDdX8E1sUsdwTWAE/G+fxBQB3wkwZlpyU4XvdGYpkOrAQ6x9nWI+a9A6c22L4E+HGDMpcB\n/wS2AH8CSoHvNfjcqLDs2HC5G3A7sBooB14FijL9d9Ir+16qSUhOMrNhwPFAdczq44DewPUNy7v7\nO8CLwBkNysatMbj7xgTH7Rke92/uvjnO5zZEP4vtfkWQePYD/go8BJzZoMyZwAJ3f9fMjKDWMgA4\nCRgLvAa8ZGb9d+H40oopSUguOd7MNpvZVuBTYDTwh5jto8KfHyb4/AJgr/D9yCbKJjICsF34XGMe\ncfc73X2Ruy8G7gMOMbMRMWXOAO4P3x8NHEhQS5nl7iXufjWwiOC2mUhkShKSS14juDiOJ/jGPR24\nqRmfN4JbNvXvd8Wufq4xxbEL7j4X+ICw1mNmhwDDgQfDIuMIb5eFSXOzmW0GxoTlRCJTkpBcUhF+\na/7A3b9PcKG8Omb7x+HP0Qk+vw/wSYOy+zQzhk8IEk2Uzzk7J5W2ccptibPuAb645XQm8Lq7Lw2X\n84DPCRJm7Gtvdvx9iDRJSUJy2a+Bn5nZnuHyfwh6Nf2kYUEzOwg4huDiW192LXBlvB2bWfd46919\nPfAccLmZdW7ic2sIemPVb+sXu9yEB4ARZjYB+DZf3GqCoGdVP6AuTJqxr9UR9y8CKElIDnP3V4D5\nwC/C5S3ARcBXzOwuMxtrZoPNbDIwDXgDuDGm7HcI2jmeMbNjzazQzA4ys+v4IpnE812CGkKxmX3L\nzPYys73N7FJgbky5l4DLzKzIzMYS9NCqjHhupQS3124l6Mn0WMzmF4A3gafM7AQzG2pmE83s12b2\npSj7F6mnJCG57k/AhWY2BMDdnwCOAPoSXKQ/Iahx3Al82d2r6j/o7k8BE4EKgm/qCwkuxoOAnyY6\nYNi4fBDwPEHD+dzwWCcDF8cU/RFBY/IrwONhDM35pn8fcADwTGxvK3d34MTwmHeEcT9K0Ci/shn7\nF8GCf08iIiI7U01CREQSUpIQEZGElCRERCQhJQkREUmoTaYD2F29e/f2wsLCTIchIpJV5syZs9bd\n+zRVLuuTRGFhIcXFxU0XFBGR7cxsadOldLtJREQaoSQhIiIJKUmIiEhCShIiIpKQkoSIiCSUtiQR\njrq52szmJdhuZnaTmZWY2dxw6GYREcmgdNYkphLM/ZvICQRTRo4EpgC3pCEmERFpRNqek3D318ys\nsJEipwD3hsMczzCz7mbW391XpSVAEclp7k5tnVNT59R58LO2duflWndq6+qoqXNqamO2xbyC5Tpq\n69hedsdtOy7XxXxmx+UdP9OwbK2H+68Ny9afQxjnocN78cNJo5o++d3Qkh6mGwAsj1kuDdftlCTM\nbApBbYPBgwenJTiRbJLoglgbs77hBbE5F7n42+Jc5Bq7iPoXF+mGF8TYC3NNg7ibuuDGP4c66nJw\nVoQ+Xdql/BgtKUnEm0A+7p/V3W8HbgcoKirKwT+9tDTVtXU8N/8zNlfWxL+Iengxq78gxlwAE35r\nrauLe1FL/K01yoX5iwuq7Cw/z8g3Iz/PaJNn5OfvuJxXvz7PaJOXt8Nyfl7M5+K8z8/LC/ZhX+y7\n4XJ+/fu8PNrkx2xLuM8gjvw8tu8/dlvvzgUp/521pCRRSjDjV72BaBYtybC6Ouflhav5438+ZsGq\nTZkOp1nyjPACs/MFsNGL3A4Xyh3L5pnRJj/+BTHexbf+gpjoIhfvwhxbtmEsDc+hTYK48+2Li3Rs\nYjCL911UGtOSksQ0gsnjHwYOAcrUHiGZUl1bx7T3VnLba5/y8eebAWiTZxy1V1/6dm234wWxwQUr\nKRe5OBfm7fuPuQAm/NaqC6IkSdqShJk9BBwF9DazUuBXQFsAd78VmE4wL28JwZzC56crNpF6W7bV\n8NCsZdz1xmJWllUCsEfX9nznS0OZPH4wndu1pO9VIqmXzt5Npzex3YHL0hSOyA7Wbt7GPW8t4d63\nl1K2tRqAEX07c/ERwzjlwAEUtNFzp9I66WuRtGrL1lVwx+uLeLR4Odtq6gAYN6QHlxw5nGP27kte\nnm7ZSOumJCGt0rwVZdz66qdM/2DV9q6Rk/bpyyVHDqeosGdmgxNpQZQkpNVwd976dB23vvopr3+y\nFggao78+dgAXHzmMUf26ZDhCkZZHSUJyXm2d8+y8Vdz26iI+WFEGQMeCfE4fP5gLDx/Knt07ZDhC\nkZZLSUJyVmV1LY/PKeWO1xexdF0FAL06FXD+YYWcNWEI3Tum/kEkkWynJCE5p6yimvtnLuXuNxez\ndnMVAIN7duSiI4bxrXEDad82P8MRimQPJQnJGavKtnLXG4t5cOYytlTVArDvnl255MjhnDBmD9rk\nqxurSHMpSUjWK1ldzq2vLuKp91ZQXRt0VTp8RG8uPnIYh4/orSePRXaDkoRkrTlL13PLK4t44cPP\ngWCsoq/s359LjhjOfgO7ZTg6kdygJCFZZ+Fn5Zxw42vbn28oaJPHt8YN5KIvDaOwd6fMBieSY5Qk\nJKvMX1nGqbe8vT1BXH70CM49tDAt4+qLtEZKEpI1pn+wip89Ppet1bV8eXQ/rj91f3VjFUkxJQnJ\nCne9sZjrnlmAOxy3bz9uOn0s7dqoK6tIqilJSItWV+f84d8fcdtriwC48oS9ufiIYeqxJJImShLS\nYlXV1PHTx9/nyfdW0ibPuP7U/fnGQQMzHZZIq6IkIS3S5m01XHLfHN4oWUvHgnxuPWscR4zqk+mw\nRFodJQlpcVaXV3L+3bOZv3ITvTsXcPd54/Xcg0iGKElIi7JozWbOuWsWpRu2UtirI/decAiDe3XM\ndFgirZaShLQY7y7bwAVTZ7OhopoDBnbjrvMOpldnPf8gkklKEtIivPjh51z24DtUVtdx9F59uPnM\ng+hYoH+eIpmm/4WScQ/PWsbPn5xHbZ1zWtFAfvv1/WirEVtFWgQlCcmoB2Yu5edPzAPge/81giuO\nHaVnIERaECUJyQh3Z+pbS7juXwsA+J8T92bKEcMzHJWINKQkIWlXUVXDjx97n+kffAbAD44ZyUVf\nGpbhqEQkHiUJSauHZi3jphc/YVVZJQA3nLo/3yoalOGoRCQRJQlJi4qqGn7y2Fye+WAVAAO6d+CP\npx3AhGG9MhyZiDRGSUJSrqyimnPvnsV7yzdS0CaPn3x5L84/rFBzTotkASUJSamZi9ZxxaPvs2Lj\nVgZ078A9FxzMiL5dMh2WiESkJCEpsbq8kvvfXspNL5UAMGZAV245cxyDemqIDZFsoiQhSVVVU8ej\nxcu54bmFlG2tBmDywYO47mtj9ICcSBZSkpCkqK6t47HiUm5+uYQVG7cCMG5ID3507CgmDu+lB+RE\nslTkJGFmbYHjgOHA3e6+ycwGAWXuviniPo4HbgTygTvd/fcNtg8G7gG6h2WudPfpUWOUzKisruX8\nu2fz9qJ1AIzs25kfTBrJiWP6k5en5CCSzSIlCTMrBJ4H+gEdgaeBTcCPgA7AxRH2kQ/cDBwLlAKz\nzWyauy+IKfYL4FF3v8XMRgPTgcKI5yIZUFldyzl3zWLW4vX06lTANSfvy1f2U3IQyRVRbxLfCLwJ\n9AK2xqx/Ajgm4j7GAyXuvsjdq4CHgVMalHGga/i+G7Ay4r4lA1Zu3Mql989h1uL1FOTncePksXz1\ngD2VIERySNTbTYcBh7l7dYN7y0uBPSPuYwCwPGa5FDikQZlrgP+Y2feATsCkeDsysynAFIDBgwdH\nPLwk0z/fKeWqf37Atpo6urRrw0NTJjBmgGaPE8k1UWsS+eGroYFAecR9xPt66Q2WTwemuvtA4ETg\nPjPbKUZ3v93di9y9qE8fzXucTqs3VTLl3mKuePR9ttXUcejwXjz9vcOVIERyVNSaxPPA94BLw2U3\ns07Ar4B/R9xHKRA7SM9Adr6ddCFwPIC7v21m7YHewOqIx5AUKlm9mQumzmbZ+go6FuTz0+P24pyJ\nhbq9JJLDoiaJHwOvmNlcoD1wLzCKoBZxdsR9zAZGmtlQYAUwGTijQZllBG0cU81sn/BYayLuX1LE\n3XmsuJSf/t9cAHp3bscjF09geJ/OGY5MRFItUpJw92Vmtj9BQhhHcJvqEeAed490u8nda8zscuA5\ngltXd7n7fDO7Fih292kEvaXuMLP/JrgVdZ67N7wlJWm0YUsVv3hqHs/MDQbmGz+0J3eeW0TX9m0z\nHJmIpINFuQab2XhgjrvXNlifD4xz91kpiq9JRUVFXlxcnKnD57SXF67mZ4/PZXX5NjoW5HP1SaOZ\nfPAgPRgnkgPMbI67FzVVLurtpreB/uzcNtA93BavUVuy1OZtNVwzbT6PzykFoGhID/542gEM6dUp\nw5GJSLpFTRLGzj2RAHoAFckLRzJta1UtZ905k/eWbwTgsqOHc8Wxe5GvxmmRVqnRJGFmj4ZvHbjT\nzLbFbM4HDgBmpCg2SbOyrdWc84+ZvF9aRs9OBdx29jgOLuyZ6bBEJIOaqknUt0EYUBezDMGT1w8A\nt6QgLsmAG577iPdLy9izW3vuPn88e+2heR9EWrtGk4S7nw5gZkuA37j7lnQEJen3ysLVPDBzGfl5\nxm1nFylBiAgQvQvsVakORDLn/eUb+e4D7+AOlxw1jP0G6ulpEQk0Z6jw0wmGzRgMFMRuc/fRSY5L\n0mTx2i2ce/csKqpq+cbYAfzo2L0yHZKItCCRxm4ysx8CtwKfAnsDLxEM1rcn8HjKopOUqqyuZcq9\nxWysqOaovfrwh1P31xAbIrKDqAP8XQpMcff/BqqBP7n7ccBNgEbYy0JL121h4u9e5JPVm+nZqYC/\nnj5W04uKyE6iXhUG8UVX161AfavmfcBpyQ5KUqu6to6fPj6XDRXBHNS3nT2OLhpmQ0TiiJokPgfq\nO8wvI5hACGAI8YcAlxbsN/9awMzF6+nduYDXf3q0noUQkYSiJomXgZPC9/cAfzGzZ4FHgadSEZgk\nn7vz+2c/4p63l9Imz7jt7HEM6tkx02GJSAsWtXfTJfVl3f2vZraJYLa6F4G/pig2SaKyimque2YB\nj88pJc/gt18fw7ghqkGISOOiPidRBVTFLN9DUKOQLLB+SxVn3DGDjz4rp22+ccuZ45g0ul+mwxKR\nLLBb3VnM7CQzeydZwUjybayo4sw7Z/LRZ+UM79OJp793uBKEiETWZJIws7PN7D4zu8vMDgrXTTCz\nGcD/AR+kOkjZNTW1dZw/dTYfrtrEsN6deOiiCey9R9dMhyUiWaTRJGFmPwDuAsYSPG39arhuOkFj\n9lB3PzflUcouufvNJby7bCN7dmvPgxdNoG/X9pkOSUSyTFNtElOAy939NjM7lmDq0W8Ao9x9bcqj\nk102e8l6/vj8QgB++/X92KObEoSINF9TSaIQ+DeAuz9vZjXAlUoQLdu8FWWcdtvbuMMpB+7J0Xv3\nzXRIIpKlmmqT6EDwhHW9bQQP1kkLtbq8kvPunoU7TBzWiz98c/9MhyQiWSxKF9jzzGxzTPmzzGyH\nmoS7/z3pkckuufmlEtZurmL80J7ceW4R7dtq+nER2XVNJYnVwH/HLG8kGOwvlgNKEi3AsnUVPDRr\nOWZw3Slj6NQu8kjwIiJxNTUz3R7pCkR2T22d8+PH36eqto5vjB2gmeVEJCk0NnSOuGbafGYtXk+v\nTgX85HhNHCQiyaEkkQOWravg/plLAbhx8lj6d+uQ4YhEJFcoSWS5yupazpsa9GaatE8/Dh/ZO9Mh\niUgOUZLIcg/OXMaiNVsY3qcTv//mfpkOR0RyjJJEFttaVcvNL5cAcOUJ+9C7c7sMRyQiuSZykjCz\ntuGorz8ws67hukH17yW9yrZWc+QNL7NuSxV79evCpH30VLWIJF+kjvRmVgg8D/QDOgJPA5uAHxE8\nlX1xasKTRP7ywsesLt9Gh7b5XPHlUZhpFlkRSb6oNYkbgTeBXuw4TMcTwDFRD2Zmx5vZQjMrMbMr\nE5Q5zcwWmNl8M3sw6r5bkzlLN3D3m0vIM3jskokct68eZxGR1Ij6SO5hwGHuXt3gG+tSYM8oOzCz\nfOBm4FigFJhtZtPcfUFMmZHAVeGxNpiZ7qHEcf2/PwLgwsOHMmZAtwxHIyK5LGpNIj98NTQQKI+4\nj/FAibsvCqdDfRg4pUGZi4Cb3X0DgLuvjrjvVuO95RuZuXg9ndu14fvHjMx0OCKS46ImieeB78Us\nu5l1An5FOJR4BAOA5THLpeG6WKOAUWb2ppnNMLPj4+3IzKaYWbGZFa9Zsybi4XPD38PeTGdNGEKX\n9m0zHI2I5Lqot5t+DLxiZnOB9sC9BBf0cuDsiPuI17LqceIZCRxFUEt53czGuPvGHT7kfjtwO0BR\nUVHDfeSsjz7bxH8WfE5BmzwuOLww0+GISCsQKUm4+zIz2x84BziIoAbyCHCPu0e93VQKDIpZHgis\njFNmhrtXA4vNbCFB0pgd8Rg5q6Kqhu8+8A4A3y4aRN8ummlORFIvahfYru6+id0bEnw2MNLMhgIr\ngMnAGQ3KPEkwl/ZUM+tNUFtZtBvHzBm/eGLe9ierf3bC3pkOR0RaiahtEp+b2WNmdrKZ7dIkBe5e\nA1xOME/2h8Cj7j7fzK41s5PDYs8B68xsAfAy8BN3X7crx8slH5SW8cR7KyjIz+OWs8bRWfNEiEia\nRL3afJvgW/9DwFYzewy4z93fas7B3H06ML3Bul/GvHfgivAlgLvzy2nzcIdzDx3CqH6aJ0JE0idS\nTcLdp7n7ZIInrn8EDANeNbNFZnZtKgNs7f7xxmLeXbaRnp0K+OGkUZkOR0RamWYN8Ofum939Hnc/\nDtgfKAN+npLIhIqqGv7ywicAXHPyvpqOVETSrllJwszamdmpZvYE8C7QG/h/KYlMmP7BZ2zeVsPY\nwd05+YBID7aLiCRV1N5NxwBnAt8IV/0TOBF4OWxHkCSrq3PueWsJEHR5FRHJhKj3L6YT9DyaAjzl\n7ttSF5IAPFK8nA9WlNGvazu+qlqEiGRI1CTR393XpzQS2W5u6Ub+95kPAfjFV0arLUJEMibh1cfM\nOrp7RbhYaWYdE5WNKSe7yd35xZPzKN9Ww2EjenHS/v0zHZKItGKNfUUtN7P+4Uism9l5nKVY8UaI\nlV3wysdrmFtaRseCfG4+4yBNJiQiGdVYkjgRWB/zXg3UaVDfWP39Y0bSvWNBZoMRkVYvYZJw9+di\n3kcdDlx2w8LPynn14zUU5Odxmno0iUgLEOk5CTOrMLM+cdb3NDO1RyTJz5/4AHc47eCB9OykWoSI\nZF7Uh+naE38+iPbN2Ic0YnV5JcVLN9ChbT4/O16jvIpIy9Bo30oz+2741oHzzGxzzOZ84Ejg4xTF\n1qq8+GEwU+vYwd0145yItBhNdcC/OvxpBAP71cVsqwKWAN9FdkvJ6s38+un5AJwwZo8MRyMi8oVG\nk4S79wcws7eBE919Q1qiamV+/fR8Kqvr+PLofpxxyJBMhyMisl3U6UsnpjqQ1mp1eSUzFwc9jX/z\ntTHk5+m5CBFpORp74vp64NfuviV8n5C7/zTpkbUSt726iKqaOo7btx99u2reahFpWRqrSXwJaBvz\nPhE9ZLeLyiqquX/GUiB4eE5EpKVp7GG6ifHeS/K8+elattXUMX5oT/bds1umwxER2ckuP+NgZgPN\nTMOT7qLK6lr+/HzQe/jovfpmOBoRkfiiPnF9jZmdFbP8L2AZ8JmZFaUquFx2/4ylfLJ6MwO6d+Cs\nCYMzHY6ISFxRaxLnAZ8CmNlxwETgKOAx4PepCCzXPTvvMwCuPmkfPTwnIi1W1NtFewCl4fsTgcfc\n/TUzWwXMSklkOay2zvn483IADhzUI8PRiIgkFrUmsR4YGL4/DngxfG9oLolmm7+yjPLKGgb37Mge\n3dTtVURarqg1iSeB+83sQ6AvUD90+IFASSoCy2VvfboOgEOH98pwJCIijYtak/ghcBewAjje3cvD\n9UOAO1MRWC6bsShIEocM65nhSEREGhd1WI4q4Ldx1t+Q9IhyXHllNW+WrAVgvwHdMxyNiEjjIj/n\nYGY9gUuA0QRPWc8Hbnf39Y1+UHbw9PurqK51hvXuxPA+nTIdjohIo6I+J3EIQRfYS4B2BJMNfRco\nMbODUxdebqmtc/7yQvAA3aVHDcdMg/mJSMsWtSbxR4LG64vcvQYgfNr6TuDPwOGpCS+3vPXpWlaX\nb2NA9w5886CBTX9ARCTDoiaJccB36hMEgLvXhKPDFqckshz03PzgAbpvHjSAPA0JLiJZIGrvpnJg\nUJz1A8NtkZjZ8Wa20MxKzOzKRsqdamaeS0N+VNfWbZ+idNLofhmORkQkmqhJ4lHgH2b2TTPrb2Z7\nmNmpwB3htiaZWT5wM3ACQeP36WY2Ok65LsD3gZkRY8sKf33xE1aVVVLYqyNjNOKriGSJqLebfkww\nt8TDfJFY6gjaJH4ScR/jgRJ3XwRgZg8DpwALGpS7Drg+PGZOcHcemr0cgP/9+n661SQiWSNSTcLd\nK939YqAPMIFggL8+7n6pu1dGPNYAYHnMcmm4bjszGwsMcvd/NbYjM5tiZsVmVrxmzZqIh8+cBas2\nsaZ8G3t0bc9EPWUtIlmkyZqEme0JHENQk3jN3Wfv4rHifX3ePqudmeUR9JQ6r6kdufvtwO0ARUVF\nLX5mvFc/DhLZkaP6qNuriGSVRpOEmR0KTAe6hquqzOwsd398F45Vyo6N3wOBlTHLXYAxwCvhhXQP\nYJqZnezuWd2D6tWFYZLYq0+GIxERaZ6mbjf9BpgBjCC4qD8I/L9dPNZsYKSZDTWzAmAyMK1+o7uX\nuXtvdy9098LwuFmfIEpWlzNz8Xry84zDRvTOdDgiIs3S1O2mA4CjYxqbfwBsNLPu7r6xOQcKn6u4\nHHiOYHjxu9x9vpldCxS7+7TG95Cdrnj0fQAmDutFtw6aXEhEsktTSaIH8Fn9gruXm1lFuL5ZSSL8\n/HSC21ex636ZoOxRzd1/S7Opspq5pWXkGfxl8oGZDkdEpNmidIEdZWax90mM4LZRh/oV7t6wG6sA\nT78fNLkcNLgHvTu3y3A0IiLNFyVJvNpg2QgmHfLwvaPZ6eKasSgYIPcbGqdJRLJUU0lin7REkaNW\nbKgAYJiGBBeRLNVoknD3hekKJNe4O5+u2QLAsN5KEiKSnaKO3STNtLp8G2Vbq+nSvg19uqg9QkSy\nk5JEivx7XtApbNyQHnrKWkSylpJEirz+STCPddGQHhmORERk1ylJpMCmympmLwl6Nh29d98MRyMi\nsuualSTMrLOZHWBmenS4ETe/VELZ1moOLuzB6P5dm/6AiEgLFSlJmFknM7sX2ATMIRyoz8z+ZmY/\nT2F8WaeiqobbXlsEwC++MlrtESKS1aLWJH4H7A0cCsTOH/Ef4FvJDiqbvbcsGK2kXZs8DhjUPcPR\niIjsnqgz050CnObuM80sdv6GBcCw5IeVvWYsDtoiTiuKNyW4iEh2iVqT6AOsjrNeT4nFqK6t44EZ\nSwE0LLiI5ISoSWIOcGLMcn1t4gLg7aRGlMXe+GQt67ZUMbxPJ47bt1+mwxER2W1Rbzf9HJhuZnuH\nn7nMzPYFjgKOTFFsWeeR2cEU3l8fO0AN1iKSEyLVJNz9NYJk0BdYAXwD2AIc5u6zUhde9liwchP/\nnv8ZBW3y+NrYAZkOR0QkKaLWJHD3OcC3UxhLVrt/ZtAWcdJ+/RnYo2OGoxERSY5IScLMGr3quXtF\ncsLJXnOWbABg7GB1exWR3BG1JrGZLxqr42n1kw6VhnNHnLBf/wxHIiKSPFGTxAkNltsCY4HvAFcn\nNaIs9Mnn5WypqqVD23x6dizIdDgiIkkTKUm4+3NxVv/LzD4GzgLuTWpUWWbJuqAWceCg7uTlqVeT\niOSO3R0Fthj4r2QEks2e/WAVAKP6dc5wJCIiybXLScLMCoDLCLrEtlpL123hqfdXkp9nnH/Y0EyH\nIyKSVFF7N61hx4ZrA7oDVcA5KYgra9zw3EJq65xTxw2kUHNZi0iOidpw/YsGy3XAGuAtd483plOr\nMWdp0PX1kiOHZzgSEZHkazJJmFkboBqY7u6fpT6k7LGqbCuryipp1yaPgT06ZDocEZGka7JNwt1r\ngL8B7VIfTnaZt2ITAOOG9KB921b/qIiI5KCoDdezgANSGUg2+nBVkCRG9euS4UhERFIjapvE34A/\nmtmeBMOGb4nd6O4Lkh1YNvj483IARu+peaxFJDdFTRKPhj//Hv6s7+lk4ftWd6+loqqGd8JGa9Uk\nRCRXRU0S+yTjYGZ2PHAjQVK5091/32D7FQRDfdQQ9J66wN2XJuPYyfbI7OWsLKtkWO9OjO6vmoSI\n5KZGk4SZ3QX8wN0X7u6BzCwfuBk4FigFZpvZtAa3qt4Fity9wswuBa6nhQ5PPntJMJf1lCOGUdBm\ndx9cFxFpmZq6up0LJKtv53igxN0XuXsV8DBwSmwBd385ZtjxGcDAJB076VZv2gagB+hEJKc1lSSS\nOVrdAGB5zHJpuC6RC4Fn4wZlNsXMis2seM2aNUkMMbrPNlUC0KuTRn0VkdwV5T5JY/NINEe8hBN3\n32Z2FlBJC+UnAAAO8UlEQVQE3BA3IPfb3b3I3Yv69OmTpPCiW7J2C6UbttKpIJ+hqkmISA6L0nD9\nmVnjFQp3j9K7qRQYFLM8EFjZsJCZTQJ+Dhzp7tsi7DftnpsfPHh+3L570CZf7REikruiJIkpwMYk\nHGs2MNLMhhKMHDsZOCO2gJmNBW4Djm/JY0K9Xxr8OiYM65XhSEREUitKkng6GRdsd68xs8uB5wi6\nwN7l7vPN7Fqg2N2nEdxe6gw8FtZelrn7ybt77GQqWV3OCwuCX8dBQ3pkOBoRkdRqKkkkqz0i2Jn7\ndGB6g3W/jHk/KZnHS4U7X19MVW0d3y4axIi+mmRIRHJbOns3Zb2a2jqenRe0R5xz6JAMRyMiknqN\n1iTcXa2yMcq2VlO2tZou7dqw757dMh2OiEjKKQk0wyerNwPQu4tGTReR1kFJohnufXsJAF/Zr39G\n4xARSRcliYhWbNzKc/M/p02ecdYEtUeISOugJBHRfW8vpbbOOWG//uzRrX2mwxERSQsliQi2VtXy\n8OxlAJx/WGFmgxERSSMliQiefG8FGyuqOWBgN8YO6p7pcERE0kZJognuztQ3lwBw/mFDaWocKxGR\nXKIk0YS3P13Hws/L6dOlHSeqV5OItDJKEk24+60lAJx1yBDNQCcirY6ueo1Ytq6CFz78nIL8PM44\nZHCmwxERSTsliUbc+/YS3OGkA/rTR09Zi0grpCSRwJZtNTxSHMy2ev6hQzMcjYhIZihJJPDPd0op\nr6yhaEgP9huowfxEpHVSkoijrs63N1ifp4fnRKQVU5KI4/WStSxas4X+3dpz3L57ZDocEZGMUZKI\n4+43FwNw1oQhtM3Xr0hEWi9dARtYtGYzryxcQ7s2eZwxXt1eRaR1U5Jo4J6wLeLrYwfQo1NBZoMR\nEckwJYkYmyqreXxOKaAGaxERUJLYwWPFpWypqmXisF7svUfXTIcjIpJxShKh2jrffqtJtQgRkYCS\nROjlj1azbH0FA3t0YNI+/TIdjohIi6AkEZoa1iLOnVhIfp7mjBARASUJAD7+vJw3StbSoW0+px08\nKNPhiIi0GEoSwN3hzHPfHDeAbh3aZjYYEZEWpNUniY0VVTzxbtjt9dDCzAYjItLCtPok8fDs5VRW\n1/Glkb0Z0bdLpsMREWlRWnWSqKmt4763lwJwwWGaM0JEpKFWnSSeX/A5KzZuZWjvThw5qk+mwxER\naXHSmiTM7HgzW2hmJWZ2ZZzt7czskXD7TDMrTGU8d2/v9jqEPHV7FRHZSdqShJnlAzcDJwCjgdPN\nbHSDYhcCG9x9BPBn4A+pimf+yjJmLV5P53Zt+Oa4gak6jIhIVktnTWI8UOLui9y9CngYOKVBmVOA\ne8L3jwPHmFlKvuJPDbu9fqtoIF3aq9uriEg86UwSA4DlMcul4bq4Zdy9BigDejXckZlNMbNiMyte\ns2ZNswNxd2rdaZtvnDuxsNmfFxFpLdKZJOLVCHwXyuDut7t7kbsX9enT/AZnM+NPpx3IrP+ZRGHv\nTs3+vIhIa5HOJFEKxI55MRBYmaiMmbUBugHrUxWQJhUSEWlcOpPEbGCkmQ01swJgMjCtQZlpwLnh\n+1OBl9x9p5qEiIikR5t0Hcjda8zscuA5IB+4y93nm9m1QLG7TwP+AdxnZiUENYjJ6YpPRER2lrYk\nAeDu04HpDdb9MuZ9JfCtdMYkIiKJteonrkVEpHFKEiIikpCShIiIJKQkISIiCVm29zA1szXA0l38\neG9gbRLDyQY659ZB59w67M45D3H3Jp9GzvoksTvMrNjdizIdRzrpnFsHnXPrkI5z1u0mERFJSElC\nREQSau1J4vZMB5ABOufWQefcOqT8nFt1m4SIiDSutdckRESkEUoSIiKSUKtIEmZ2vJktNLMSM7sy\nzvZ2ZvZIuH2mmRWmP8rkinDOV5jZAjOba2YvmtmQTMSZTE2dc0y5U83MzSzru0tGOWczOy38W883\nswfTHWOyRfi3PdjMXjazd8N/3ydmIs5kMbO7zGy1mc1LsN3M7Kbw9zHXzA5KagDuntMvgmHJPwWG\nAQXA+8DoBmW+C9wavp8MPJLpuNNwzkcDHcP3l7aGcw7LdQFeA2YARZmOOw1/55HAu0CPcLlvpuNO\nwznfDlwavh8NLMl03Lt5zkcABwHzEmw/EXiWYGbPCcDMZB6/NdQkxgMl7r7I3auAh4FTGpQ5Bbgn\nfP84cIyZxZtKNVs0ec7u/rK7V4SLMwhmCsxmUf7OANcB1wOV6QwuRaKc80XAze6+AcDdV6c5xmSL\ncs4OdA3fd2PnGTCziru/RuMzdJ4C3OuBGUB3M+ufrOO3hiQxAFges1warotbxt1rgDKgV1qiS40o\n5xzrQoJvItmsyXM2s7HAIHf/VzoDS6Eof+dRwCgze9PMZpjZ8WmLLjWinPM1wFlmVkowf8330hNa\nxjT3/3uzpHXSoQyJVyNo2O83SplsEvl8zOwsoAg4MqURpV6j52xmecCfgfPSFVAaRPk7tyG45XQU\nQW3xdTMb4+4bUxxbqkQ559OBqe7+RzObSDDb5Rh3r0t9eBmR0utXa6hJlAKDYpYHsnP1c3sZM2tD\nUEVtrHrX0kU5Z8xsEvBz4GR335am2FKlqXPuAowBXjGzJQT3bqdleeN11H/bT7l7tbsvBhYSJI1s\nFeWcLwQeBXD3t4H2BAPh5apI/993VWtIErOBkWY21MwKCBqmpzUoMw04N3x/KvCShy1CWarJcw5v\nvdxGkCCy/T41NHHO7l7m7r3dvdDdCwnaYU529+LMhJsUUf5tP0nQSQEz601w+2lRWqNMrijnvAw4\nBsDM9iFIEmvSGmV6TQPOCXs5TQDK3H1Vsnae87eb3L3GzC4HniPoGXGXu883s2uBYnefBvyDoEpa\nQlCDmJy5iHdfxHO+AegMPBa20S9z95MzFvRuinjOOSXiOT8HfNnMFgC1wE/cfV3mot49Ec/5R8Ad\nZvbfBLddzsvmL31m9hDB7cLeYTvLr4C2AO5+K0G7y4lACVABnJ/U42fx705ERFKsNdxuEhGRXaQk\nISIiCSlJiIhIQkoSIiKSkJKEiIgkpCQhLZKZtQlHav1apmPZVWY2IjyHA5sod7+ZPZmuuESaQ0lC\nUsLMpoYXyIavRi+Y6WRmv4mJq9bMlpnZ7WaWrHG7FgP9gXnh8SaFx+reoNxlpHi4kJhj17/WhUPE\nT2jmfrI+eUvzKElIKr1AcJGMfcUdEz+D5hPENRi4HPg6MDUZO3b3Wnf/LBw0srFyZWkcS2kvgvM9\nGtgAPBs+iS0Sl5KEpNK28CIZ+6oBMLMTzewNM9toZuvN7Fkz2yvRjsIhB64xs6Vmts3MVpnZ3THb\n88zsKjNbZGZbzewDMzs9Qow1YVwrwqd1/wacYGbtwv0eYGYvhftcZ8EEMPXDUMdu32Rm5Wb2npkd\nGW7bfrvJzEYAz4cf2xCuvzMst/12k5ldZmYrwwEJY8//UTP7v5jlU8zsHTOrDM/5unCYiqasDs93\nLvBboDtwcMx+DzGz581srZmVmdnrZjY+5vNLwp9PhOdQkoSYpAVTkpBM6QT8ieACdTTBcAJPm1nb\nBOVPA34IXEIwQN3JBOP41PsdcA7BBEqjgT8A/7DmD429leD/Rb6ZdSYY/mEDwTwG3ySYAOaOmPIP\nEwzTPB4YC1xL/LkqFofnAF98m78iTrmHCQaj+6/6FWFS+ipwf7h8InAvcBOwL/AdgqFkro16kmbW\niS9ucVXHbOpCMLfKlwgGQfyAoLbRI9xen1DOD89hQrJikhYqEzMt6ZX7L4JbNjXA5pjXs42U7wrU\nARPC5TYE4+58LVz+KbAAaBPns10ILswTG6z/GzCtkWP+BngvZnkfglnP3gyXLyUYy6tTTJlJYVxD\nw+UtwJkJ9j8iLHtgg892b1DufuDJmOWngbtjls8L4ygIl98Crmqwj1MJBnZLdK71x67/W3j4mhnv\ndxrzOSMYHG9yvL9LTLlmx6RXdrxUk5BUeg04MOb1nfoNZjbSzB4Kb0tsIhja2AjaBuJ5hCAZLDaz\nOy2Yp7r+VsYYoB3wvJltrn8RzMo2vIkY9wvLbyVon1gCnB1u2wd43923xJR/M2YbBLWhqWb2gpn9\nj5mNauJ4UdwPfMPM2ofLZwKPeTATG8A44JcNzvVeoKuZ9Wli318imArzdILazTke02ZiZv3CxvuP\nzawMKCeYgCvR36Xe7sQkLVjOjwIrGVXh7iUJtj1DcJG6iCBB1BHUFOLew3b3peEFeBLBMNB/Bq62\nYFKZ+i87XwFWNPhoFY1bSHDrqhZY6TvOq2HsPHmLx/5096vN7D6CUTi/DFxjZhe5+z3suqcI5mn+\nqpm9QXA7Lva2jRGMBPrPOJ9tah6UxR40kn8c3nJ6wswOcPf6W073E7RT/BBYCmwDXiHB3yVJMUkL\npiQhaWdm/QjaFS5099fDdeNpoo3M3bcS3Ip52sxuIJhsZQIwhyAZDHb3V5sZTlUjiWwBcKaZdYqp\nTRwe/vwwJq6PgY+Bv5jZHQST3sRLEvUJK7+xgNy90sz+SVCDGEhwnm/EFHkX2KuRuKOaClxNcFvt\npnDd4cAUd58OYMFcyXvEfKY2fDU8h2TFJC2MkoRkwlqCb5dTzGwVwYXwBoLaRFxmdkH4dhZBO8AZ\nBA2uJe5eZmZ/Bv5sZvnA6wRtHBMJksCduxjnfQTfju8xs2sIGpRvAR519yVhw/bvgMcJblP1Bw4j\nuM0Wz9Lw51fM7Flgq7tvTlD2foJ5AvYCHnD32BrNr4GnzGw58BjBRXs/YJy7Xxn15Ny91sxuBK4y\nszvdvYIg2Z1tZsUEt/duIKhN1H/GzWwZcIyZvUnQg21DsmKSlkdtEpJ27l4LfJvg3vg84K/AVezY\ny6ahjcAUgm/UHwCnEDSeLgu3X0XQEP0zgm/5/wG+RnBLa1fj3AwcB/Qk6En1T4IEdFFYpIYgcdxL\ncHGt3/7jBPtbSnAxvR74HPhLI4d/GVgN7E3YqylmP9MJejsdG8Y1i6BhfxnNdyfQgeAZEQgaybsT\n1AweJJi9cHmDz1wRHnt5ePxkxyQtiCYdEhGRhFSTEBGRhJQkREQkISUJERFJSElCREQSUpIQEZGE\nlCRERCQhJQkREUlISUJERBL6/9cyCBZMAvOdAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2e1e950>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Separar Entrada e Saída\n",
    "y = df['fully_paid']\n",
    "X = df.drop(['fully_paid'], axis=1)\n",
    "\n",
    "#Parametros para a decision tree\n",
    "path = './classifiers/Decision Tree with augmentation/decisionTreeCLF'\n",
    "#Treinar modelo variando profundidade da arvore de 3 a 20, com 9 folds\n",
    "parameters = {'max_depth':range(3,20)}\n",
    "method = tree.DecisionTreeClassifier()\n",
    "idx_max = get_clf_with_oversampling(X, y, path, False, parameters, method, 'pkl') #Mudar para True, se quiser retreinar modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "kp56jFWcw-DX"
   },
   "outputs": [],
   "source": [
    "# Codigo para conseguir printar datas\n",
    "#%%bash\n",
    "!/C:/Users/Thullio/Anaconda3/Lib/site-packages | apt-get install graphviz\n",
    "!pip install --upgrade graphviz\n",
    "!/c/Users/Thullio/Anaconda3/Lib/site-packages | pip uninstall pyparsing\n",
    "!pip install -Iv https://pypi.python.org/packages/source/p/pyparsing/pyparsing-1.5.7.tar.gz\n",
    "!pip install --upgrade pydot\n",
    "!pip install pydotplus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oKucLpsAZUL9"
   },
   "outputs": [],
   "source": [
    "!pip install graphviz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "LZ84uwHkJ796"
   },
   "outputs": [],
   "source": [
    "# Draw Decision Tree\n",
    "from sklearn.externals.six import StringIO  \n",
    "from IPython.display import Image  \n",
    "from sklearn.tree import export_graphviz\n",
    "import pydotplus\n",
    "import graphviz\n",
    "import pydot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "8GPPGB0_jdRd"
   },
   "outputs": [],
   "source": [
    "with open('./classifiers/Decision Tree with augmentation/decisionTreeCLF{}.pkl'.format(idx_max), 'rb') as f:\n",
    "  decTreeClf = pickle.load(f)\n",
    "\n",
    "tree_model = decTreeClf.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Gc3OmqweZUME"
   },
   "outputs": [],
   "source": [
    "#from graphviz import Source\n",
    "#dot_data = StringIO()\n",
    "#export_graphviz(tree_model, out_file=dot_data, feature_names=X.columns, \n",
    "#                filled=True, rounded=True,\n",
    " #               special_characters=True)\n",
    "#graph = pydotplus.graph_from_dot_data(dot_data.getvalue())  \n",
    "#graph.write_pdf(\"./Resultados/dec_tree_with_oversampling.pdf\")\n",
    "#Image(graph.create_png())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "lvGqOrwVZUMI",
    "outputId": "5822381b-70de-418d-d7c7-634e6f121ac2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "O valor médio do hiperparâmetro escolhido foi: max_depth = 18.3\n"
     ]
    }
   ],
   "source": [
    "#Pegar as valores médios de cada classificador\n",
    "feature_importances = [0]*X.shape[1]\n",
    "max_depth_dt = 0\n",
    "for i in range(0,30):\n",
    "    with open('./classifiers/Decision Tree with augmentation/decisionTreeCLF{}.pkl'.format(i), 'rb') as f:\n",
    "      clf = pickle.load(f)\n",
    "    feature_importances = feature_importances+clf.best_estimator_.feature_importances_\n",
    "    max_depth_dt = max_depth_dt+clf.best_params_['max_depth']\n",
    "    \n",
    "print('O valor médio do hiperparâmetro escolhido foi: max_depth = {}'.format(max_depth_dt/30))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "a_WPeQQ0ZUML",
    "outputId": "f4333cf3-ad78-42cf-c717-627558e6e254"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x286089f0>"
      ]
     },
     "execution_count": 72,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiUAAAGDCAYAAAD0wGCFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nP6BlRsO3AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27d56490>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importances = feature_importances/30\n",
    "indices = np.argsort(importances)\n",
    "colors = ['0.7']*(indices.shape[0] - 1) + ['tab:blue']\n",
    "\n",
    "plt.figure(1, figsize=(8, 6))\n",
    "plt.title('Feature Importances')\n",
    "plt.barh(range(len(indices)), importances[indices], color=colors, align='center')\n",
    "plt.yticks(range(len(indices)), X.columns[indices])\n",
    "plt.xlabel('Relative Importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "IiKVUEtLZUMP"
   },
   "source": [
    "### 2.2) Random Forest with oversampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "s4fE7qTSZUMS",
    "outputId": "15e6c67a-d832-44ea-9173-81ce2f677dca"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Type I error (FP): 0.23889\n",
      "Mean Type II error (FN): 0.00013\n",
      "Mean Accuracy: 0.76098\n",
      "Mean ROC_AUC_Score: 0.86164\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAEbCAYAAAAvc3j1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/h/BsQbeAVNUIgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x368fc090>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "#Parametros para a decision tree\n",
    "path = './classifiers/Random Forest/randomForestCLF'\n",
    "#Treinar modelo variando profundidade da arvore de 3 a 20, o numero de arvores da random forest de 10 a 100 e com 9 folds\n",
    "parameters = {'max_depth':[5,10,15], \"n_estimators\":[10,30,50]}\n",
    "method = RandomForestClassifier()\n",
    "idx_max = get_clf_with_oversampling(X, y, path, False, parameters, method, 'sav') #Mudar para True, se quiser retreinar modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9FZR7D37ZUMV"
   },
   "outputs": [],
   "source": [
    "with open('./classifiers/Random Forest/randomForestCLF{}.sav'.format(idx_max), 'rb') as f:\n",
    "  clf = joblib.load(f)\n",
    "\n",
    "random_forest = clf.best_estimator_\n",
    "\n",
    "dot_data = StringIO()\n",
    "export_graphviz(random_forest.estimators_[5], out_file=dot_data, feature_names=X.columns, \n",
    "                filled=True, rounded=True,\n",
    "                special_characters=True)\n",
    "graph = pydotplus.graph_from_dot_data(dot_data.getvalue())  \n",
    "graph.write_pdf(\"./Resultados/random_forest_with_oversampling.pdf\")\n",
    "Image(graph.create_png())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "o_ZlnjCmZUMa",
    "outputId": "9ff14255-6ba4-4f13-c565-e004c4553326"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Os valores médios dos hiperparâmetros escolhidos foram: max_depth = 15.0 e n_estimators = 44.0\n"
     ]
    }
   ],
   "source": [
    "#Pegar as valores médios de cada classificador\n",
    "feature_importances = [0]*X.shape[1]\n",
    "max_depth_rf = 0\n",
    "n_estimators = 0\n",
    "for i in range(0,30):\n",
    "    with open('./classifiers/Random Forest/randomForestCLF{}.sav'.format(i), 'rb') as f:\n",
    "      clf = joblib.load(f)\n",
    "    feature_importances = feature_importances+clf.best_estimator_.feature_importances_\n",
    "    max_depth_rf = max_depth_rf+clf.best_params_['max_depth']\n",
    "    n_estimators = n_estimators+clf.best_params_['n_estimators']\n",
    "    \n",
    "print('Os valores médios dos hiperparâmetros escolhidos foram: max_depth = {} e n_estimators = {}'.format(max_depth_rf/30,n_estimators/30))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "1s074B1fZUMg",
    "outputId": "5c4c9828-baf2-46ff-e0ae-8a995a263bf5"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x274a8050>"
      ]
     },
     "execution_count": 74,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiUAAAGDCAYAAAD0wGCFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27d0b5b0>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importances = feature_importances/30\n",
    "indices = np.argsort(importances)\n",
    "colors = ['0.7']*(indices.shape[0] - 1) + ['tab:blue']\n",
    "\n",
    "plt.figure(1, figsize=(8, 6))\n",
    "plt.title('Feature Importances')\n",
    "plt.barh(range(len(indices)), importances[indices], color=colors, align='center')\n",
    "plt.yticks(range(len(indices)), X.columns[indices])\n",
    "plt.xlabel('Relative Importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "mCmH2ZxCZUMm"
   },
   "source": [
    "### 2.3) Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "aU35f3sbZUMu"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "FfB52FresBeS"
   },
   "outputs": [],
   "source": [
    "# Helper class to perform K-Folds Validation splitting\n",
    "class CrossValidationFolds(object):\n",
    "    \n",
    "    def __init__(self, data, labels, num_folds, shuffle=True):\n",
    "        self.data = data\n",
    "        self.labels = labels\n",
    "        self.num_folds = num_folds\n",
    "        self.current_fold = 0\n",
    "        \n",
    "        # Shuffle Dataset\n",
    "        if shuffle:\n",
    "            perm = np.random.permutation(self.data.shape[0])\n",
    "            data = data[perm]\n",
    "            labels = labels[perm]\n",
    "    \n",
    "    def split(self):\n",
    "        current = self.current_fold\n",
    "        size = int(self.data.shape[0]/self.num_folds)\n",
    "        \n",
    "        index = np.arange(self.data.shape[0])\n",
    "        lower_bound = index >= current*size\n",
    "        upper_bound = index < (current + 1)*size\n",
    "        cv_region = lower_bound*upper_bound\n",
    "\n",
    "        cv_data = self.data[cv_region]\n",
    "        train_data = self.data[~cv_region]\n",
    "        \n",
    "        cv_labels = self.labels[cv_region]\n",
    "        train_labels = self.labels[~cv_region]\n",
    "        \n",
    "        self.current_fold += 1\n",
    "        return (train_data, train_labels), (cv_data, cv_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ECRnYQT35ZpH"
   },
   "source": [
    "####2.3.1) Normalizing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "o7yCU3JNHLNJ"
   },
   "outputs": [],
   "source": [
    "# Separar Entrada e Saída\n",
    "y_data = df['fully_paid']\n",
    "X_data = df.drop(['fully_paid'], axis=1)\n",
    "X_data = (X_data-X_data.mean())/(pow(X_data.std(), 2) + 0.0001)\n",
    "X_data = (X_data-X_data.min())/(X_data.max() - X_data.min()).values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1524,
     "status": "ok",
     "timestamp": 1540289966871,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "yIWKibOaOeVT",
    "outputId": "e145818f-f8d0-409a-99f6-5f1fd4ba59ec"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# Parameters\n",
    "learning_rate = 0.01\n",
    "training_epochs = 100\n",
    "FOLDS = 9\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Dropout\n",
    "from keras.optimizers import Adam, SGD\n",
    "from keras.constraints import maxnorm\n",
    "from keras.utils import plot_model\n",
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from keras import regularizers\n",
    "from keras.models import load_model\n",
    "\n",
    "def create_baseline():\n",
    "  # create model\n",
    "  model = Sequential()\n",
    "  model.add(Dense(9, input_dim=15, kernel_initializer='normal', kernel_regularizer=regularizers.l2(0.01), activation='tanh'))\n",
    "  model.add(Dense(1, kernel_initializer='normal', activation='tanh'))\n",
    "  # Compile model                                                         , amsgrad=True\n",
    "  model.compile(loss='binary_crossentropy', optimizer=SGD(lr=learning_rate), metrics=['accuracy'])\n",
    "  return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 964
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 10883619,
     "status": "ok",
     "timestamp": 1540264854593,
     "user": {
      "displayName": "Thullio Debortoli",
      "photoUrl": "",
      "userId": "06174144415137317679"
     },
     "user_tz": 180
    },
    "id": "SS-QIn6bg6ko",
    "outputId": "96e1337b-9b4b-45ce-e382-f6979c3b6100"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Type I error (FP): 0.17934\n",
      "Mean Type II error (FN): 0.11204\n",
      "Mean Accuracy: 0.70862\n",
      "Mean ROC_AUC_Score: 0.78674\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEbCAYAAADajfNFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f437fbae320>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEVCAYAAAARjMm4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f439c63e7f0>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEVCAYAAADpbDJPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f439c133240>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Definir path\n",
    "path = '/content/drive/My Drive/UFMG/TCC2/Algoritmo/classifiers/MLP_oversampling/MLPCLF'\n",
    "\n",
    "encoder = LabelEncoder()\n",
    "encoder.fit(y_data)\n",
    "y_data = encoder.transform(y_data)\n",
    "seed = 10\n",
    "np.random.seed(seed)\n",
    "\n",
    "best_estimators = []\n",
    "best_histories = []\n",
    "best_accs = []\n",
    "errorI = []\n",
    "errorII = []\n",
    "accuracy = []\n",
    "roc_auc_score = []\n",
    "out_test = []\n",
    "out_prob = []\n",
    "\n",
    "for i in range(30):  \n",
    "  # Define entrada e saída para treinamento e teste\n",
    "  X_train, X_test, y_train, y_test = balanced_subsample(X_data, y_data)\n",
    "\n",
    "  #Oversampling\n",
    "  sm = SMOTE(random_state=42)\n",
    "  X_train, y_train = sm.fit_sample(X_train, y_train)\n",
    "\n",
    "  #Results Arrays\n",
    "  estimators = []\n",
    "  histories = []\n",
    "  last_acc = []\n",
    "  fold = 0\n",
    "  #Algoritmo utilizado para gerar o classificador:\n",
    "  kfold = StratifiedKFold(n_splits=FOLDS, shuffle=True, random_state=seed)\n",
    "  for train_index, val_index in kfold.split(X_train, y_train):\n",
    "    fold = fold+1\n",
    "    X_trainning, X_val = X_train[train_index], X_train[val_index]\n",
    "    y_trainning, y_val = y_train[train_index], y_train[val_index]\n",
    "\n",
    "    estimator = KerasClassifier(build_fn=create_baseline, epochs=training_epochs, batch_size=1500, verbose=0)\n",
    "    \n",
    "    history = estimator.fit(X_trainning, y_trainning, validation_data=(X_val, y_val))\n",
    "\n",
    "    estimators.append(estimator)\n",
    "    histories.append(history)\n",
    "    last_acc.append(history.history['val_acc'][-1])\n",
    "\n",
    "  idx_max = np.argmax(last_acc)\n",
    "  best_estimators.append(estimators[idx_max])\n",
    "  best_histories.append(histories[idx_max])\n",
    "  best_accs .append(last_acc[idx_max])\n",
    "\n",
    "  best_estimator = estimators[idx_max]\n",
    "\n",
    "  best_estimator.model.save('{}{}.h5'.format(path, i))\n",
    "  \n",
    "  #test and store 'test data'\n",
    "  #Testar modelo obtido\n",
    "  y_predict = best_estimator.predict(X_test)\n",
    "  y_difference = y_test - y_predict.reshape(y_predict.shape[0])\n",
    "  y_pred_proba = best_estimator.predict_proba(X_test)[::,1]\n",
    "  out_test.extend(y_test)\n",
    "  out_prob.extend(y_pred_proba)\n",
    "  errorI.append(float(y_difference[y_difference == -1].shape[0])/float(y_difference.shape[0]))\n",
    "  errorII.append(float(y_difference[y_difference == 1].shape[0])/float(y_difference.shape[0]))\n",
    "  accuracy.append(float(y_difference[y_difference == 0].shape[0])/float(y_difference.shape[0]))\n",
    "  roc_auc_score.append(metrics.roc_auc_score(y_test, y_pred_proba))\n",
    "\n",
    "print ('Mean Type I error (FP): {0:.5f}'.format(np.mean(errorI)))\n",
    "print ('Mean Type II error (FN): {0:.5f}'.format(np.mean(errorII)))\n",
    "print ('Mean Accuracy: {0:.5f}'.format(np.mean(accuracy)))\n",
    "print ('Mean ROC_AUC_Score: {0:.5f}\\n\\n'.format(np.mean(roc_auc_score)))\n",
    "\n",
    "idx_max = np.argmax(accuracy)\n",
    "\n",
    "\n",
    "y_test = out_test[int(idx_max*(len(out_test)/30)):int((idx_max*(len(out_test)/30))+(len(out_test)/30))]\n",
    "y_pred_proba = out_prob[int(idx_max*(len(out_test)/30)):int((idx_max*(len(out_test)/30))+(len(out_test)/30))]\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test,  y_pred_proba)\n",
    "plt.plot(fpr,tpr, linewidth=2)\n",
    "plt.xlabel(\"False Positive Rate\", fontsize=14)\n",
    "plt.ylabel(\"True Positive Rate\", fontsize=14)\n",
    "plt.title(\"ROC Curve\", fontsize=14)\n",
    "plt.show()\n",
    "\n",
    "history = best_histories[0]\n",
    "\n",
    "#  \"Accuracy\"\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'validation'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# \"Loss\"\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'validation'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "R0HI7cqRb1ZA"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = ['Árvore de decisão', 'Floresta Aleatória', 'RNA(MLP)']\n",
    "y = [0.23303,0.23889,0.17934]\n",
    "colors = ['0.7','0.7','teal']\n",
    "\n",
    "plt.figure(1, figsize=(8, 6))\n",
    "plt.title('Taxa de erro tipo I')\n",
    "plt.bar(range(len(x)), y, color=colors, align='center')\n",
    "plt.xticks(range(len(x)), x)\n",
    "\n",
    "# remove the frame of the chart\n",
    "for spine in plt.gca().spines.values():\n",
    "    spine.set_visible(False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "TCC2-Thullio - Local.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}