odev_2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.stats import poisson\n",
    "import numpy as np\n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt \n",
    "from scipy import stats\n",
    "import seaborn as sns\n",
    "import statsmodels.stats.api as sms\n",
    "import scipy.stats as st\n",
    "import statistics as ist\n",
    "import math \n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.stats import shapiro\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from scipy.stats import levene \n",
    "from scipy.stats import bartlett\n",
    "import statsmodels.formula.api as smf \n",
    " \n",
    " \n",
    "\"\"\"\n",
    "loan=kisinin hesabındaki para\n",
    "mortdue=kisinin onceden aldıgı mortgage kredisi\n",
    "value=kisinin istedigi yeni kredi miktari\n",
    "bad=sorgulanan müşterinin,iyi müşteri yada kötü müşteri oldugu bilgisidir.Eski borclarını ödeyip ödemediginden vs iyi/kötü müşteri diye etiketlenir.\n",
    "\"\"\"\n",
    "data=pd.read_csv(\"C:\\\\Users\\\\ekorkmaz\\\\Desktop\\\\EnkTelefonYedek\\\\kodluyoruz\\\\odev_hafta2\\\\hmelq.csv\")\n",
    "data=data.dropna()\n",
    "d1 = pd.DataFrame(data)\n",
    "d1.head()\n",
    "loan = data[data['bad'] == 1]['loan']\n",
    "grps=pd.unique(data.bad.values)\n",
    "my_dict={\n",
    "                'loan':'loan',\n",
    "                'mortdue':'mortdue',\n",
    "                'value':'value',\n",
    "                'yoj':'yoj',\n",
    "                'derog':'derog',\n",
    "                'delinq':'delinq',\n",
    "                'clage':'clage',\n",
    "                'ninq':'ninq',\n",
    "                'clno':'clno',\n",
    "                'debtinc':'debtinc'\n",
    "                \n",
    "         }\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      Count; \n",
      " bad        2018\n",
      "loan       2018\n",
      "mortdue    2018\n",
      "value      2018\n",
      "reason     2018\n",
      "job        2018\n",
      "yoj        2018\n",
      "derog      2018\n",
      "delinq     2018\n",
      "clage      2018\n",
      "ninq       2018\n",
      "clno       2018\n",
      "debtinc    2018\n",
      "dtype: int64\n",
      "\n",
      " Count \n",
      " col_0          count\n",
      "Bad Customer    1840\n",
      "Good Customer    178\n"
     ]
    }
   ],
   "source": [
    " #Count\n",
    "print('      Count;','\\n',data.count())#NaN olmayan kayıt sayısını döner \n",
    "my_Tab=pd.crosstab(index=data[\"bad\"],columns=\"count\")#kac müşteri iyi kac müşteri kötü onu gösterir\n",
    "my_Tab.index=[\"Bad Customer\",\"Good Customer\"]\n",
    "print('\\n','Count','\\n',my_Tab)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sum; \n",
      " Bad: 178 \n",
      "Loan: 38362700 \n",
      "Mortdue: 154048137.0 \n",
      "value: 215722676.0 \n",
      "yoj: 18170.0 \n",
      "derog: 303.0 \n",
      "delinq: 533.0 \n",
      "clage: 364541.65 \n",
      "ninq: 2094.0 \n",
      "clno: 44931.0 \n",
      "debtinc: 68825.56199999999\n"
     ]
    }
   ],
   "source": [
    "#Sum\n",
    "print('Sum;','\\n','Bad:',data.bad.sum(),'\\nLoan:',data.loan.sum(),\n",
    "               '\\nMortdue:',data.mortdue.sum(),'\\nvalue:',data.value.sum(),\n",
    "               '\\nyoj:',data.yoj.sum(),'\\nderog:',data.derog.sum(),\n",
    "               '\\ndelinq:',data.delinq.sum(),'\\nclage:',data.clage.sum(),\n",
    "               '\\nninq:',data.ninq.sum(),'\\nclno:',data.clno.sum(),'\\ndebtinc:',data.debtinc.sum()               \n",
    "               )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ortalama \n",
      " bad             0.088206\n",
      "loan        19010.257681\n",
      "mortdue     76337.035183\n",
      "value      106899.244797\n",
      "yoj             9.003964\n",
      "derog           0.150149\n",
      "delinq          0.264123\n",
      "clage         180.645020\n",
      "ninq            1.037661\n",
      "clno           22.265114\n",
      "debtinc        34.105829\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print('Ortalama','\\n',data.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Medyan \n",
      " bad            0.000\n",
      "loan       16900.000\n",
      "mortdue    67595.000\n",
      "value      94514.000\n",
      "yoj            7.000\n",
      "derog          0.000\n",
      "delinq         0.000\n",
      "clage        176.670\n",
      "ninq           0.000\n",
      "clno          21.000\n",
      "debtinc       34.949\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print('Medyan','\\n',data.median())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mod \n",
      "     bad     loan   mortdue     value   reason    job  yoj  derog  delinq  \\\n",
      "0   0.0  10600.0   44097.0   57280.0  DebtCon  Other  0.0    0.0     0.0   \n",
      "1   NaN  11000.0   45279.0   65481.0      NaN    NaN  NaN    NaN     NaN   \n",
      "2   NaN  12900.0   46406.0   85424.0      NaN    NaN  NaN    NaN     NaN   \n",
      "3   NaN  13900.0   49367.0   98088.0      NaN    NaN  NaN    NaN     NaN   \n",
      "4   NaN  15200.0   58606.0  100332.0      NaN    NaN  NaN    NaN     NaN   \n",
      "5   NaN      NaN   61213.0  102287.0      NaN    NaN  NaN    NaN     NaN   \n",
      "6   NaN      NaN   62933.0  106233.0      NaN    NaN  NaN    NaN     NaN   \n",
      "7   NaN      NaN   69697.0  114494.0      NaN    NaN  NaN    NaN     NaN   \n",
      "8   NaN      NaN   94727.0  114906.0      NaN    NaN  NaN    NaN     NaN   \n",
      "9   NaN      NaN   95351.0  116160.0      NaN    NaN  NaN    NaN     NaN   \n",
      "10  NaN      NaN  206126.0       NaN      NaN    NaN  NaN    NaN     NaN   \n",
      "\n",
      "     clage  ninq  clno  debtinc  \n",
      "0   219.09   0.0  19.0   28.172  \n",
      "1      NaN   NaN   NaN   33.691  \n",
      "2      NaN   NaN   NaN   40.206  \n",
      "3      NaN   NaN   NaN   42.930  \n",
      "4      NaN   NaN   NaN      NaN  \n",
      "5      NaN   NaN   NaN      NaN  \n",
      "6      NaN   NaN   NaN      NaN  \n",
      "7      NaN   NaN   NaN      NaN  \n",
      "8      NaN   NaN   NaN      NaN  \n",
      "9      NaN   NaN   NaN      NaN  \n",
      "10     NaN   NaN   NaN      NaN  \n",
      "\n",
      " Standart Sapma \n",
      " bad            0.283665\n",
      "loan       10768.607345\n",
      "mortdue    44476.368189\n",
      "value      54026.707096\n",
      "yoj            7.573012\n",
      "derog          0.580421\n",
      "delinq         0.789816\n",
      "clage         84.420116\n",
      "ninq           1.578414\n",
      "clno           9.457238\n",
      "debtinc        8.438411\n",
      "dtype: float64\n",
      "\n",
      " Min eleman \n",
      " bad              0\n",
      "loan          1700\n",
      "mortdue       5076\n",
      "value        23031\n",
      "reason     DebtCon\n",
      "job            Mgr\n",
      "yoj              0\n",
      "derog            0\n",
      "delinq           0\n",
      "clage         0.49\n",
      "ninq             0\n",
      "clno             0\n",
      "debtinc       4.03\n",
      "dtype: object\n",
      "\n",
      " Max eleman \n",
      " bad              1\n",
      "loan         89900\n",
      "mortdue     399412\n",
      "value       512650\n",
      "reason     HomeImp\n",
      "job           Self\n",
      "yoj             41\n",
      "derog           10\n",
      "delinq          10\n",
      "clage      1168.23\n",
      "ninq            13\n",
      "clno            64\n",
      "debtinc    144.189\n",
      "dtype: object\n"
     ]
    }
   ],
   "source": [
    "print('Mod','\\n',data.mode())\n",
    "print('\\n','Standart Sapma','\\n',data.std())\n",
    "print('\\n','Min eleman','\\n',data.min())\n",
    "print('\\n','Max eleman','\\n',data.max())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mutlak Degeri \n",
      " Bad: 153    0\n",
      "154    0\n",
      "155    0\n",
      "156    0\n",
      "157    0\n",
      "Name: bad, dtype: int64 Loan: 153    18200\n",
      "154    21700\n",
      "155    34100\n",
      "156     8400\n",
      "157    17400\n",
      "Name: loan, dtype: int64 Mortdue: 153     94727.0\n",
      "154     79240.0\n",
      "155    241931.0\n",
      "156     62989.0\n",
      "157     25859.0\n",
      "Name: mortdue, dtype: float64 value: 153    136877.0\n",
      "154     96784.0\n",
      "155     36486.0\n",
      "156     76718.0\n",
      "157     43684.0\n",
      "Name: value, dtype: float64 yoj: 153    15.0\n",
      "154     5.0\n",
      "155     1.0\n",
      "156     3.0\n",
      "157    16.0\n",
      "Name: yoj, dtype: float64 derog: 153    0.0\n",
      "154    0.0\n",
      "155    0.0\n",
      "156    0.0\n",
      "157    1.0\n",
      "Name: derog, dtype: float64 delinq: 153    0.0\n",
      "154    0.0\n",
      "155    2.0\n",
      "156    2.0\n",
      "157    0.0\n",
      "Name: delinq, dtype: float64 clage: 153    168.96\n",
      "154     64.51\n",
      "155    196.01\n",
      "156    131.47\n",
      "157     95.36\n",
      "Name: clage, dtype: float64 ninq: 153    2.0\n",
      "154    6.0\n",
      "155    3.0\n",
      "156    0.0\n",
      "157    1.0\n",
      "Name: ninq, dtype: float64 clno: 153    26.0\n",
      "154    24.0\n",
      "155    50.0\n",
      "156    22.0\n",
      "157    17.0\n",
      "Name: clno, dtype: float64 debtinc: 153    36.056\n",
      "154    38.079\n",
      "155    42.459\n",
      "156    29.200\n",
      "157    27.108\n",
      "Name: debtinc, dtype: float64\n"
     ]
    }
   ],
   "source": [
    " print('Mutlak Degeri','\\n','Bad:',abs(data.bad.head()),'Loan:',abs(data.loan.head()),\n",
    "               'Mortdue:',abs(data.mortdue.head()),'value:',abs(data.value.head()),\n",
    "               'yoj:',abs(data.yoj.head()),'derog:',abs(data.derog.head()),\n",
    "               'delinq:',abs(data.delinq.head()),'clage:',abs(data.clage.head()),\n",
    "               'ninq:',abs(data.ninq.head()),'clno:',abs(data.clno.head()),'debtinc:',abs(data.debtinc.head())               \n",
    "               )\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prod \n",
      " bad        0.0\n",
      "loan       0.0\n",
      "mortdue    inf\n",
      "value      inf\n",
      "yoj        0.0\n",
      "derog      0.0\n",
      "delinq     0.0\n",
      "clage      inf\n",
      "ninq       0.0\n",
      "clno       NaN\n",
      "debtinc    inf\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print('Prod','\\n',np.prod(data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CumSum\n",
      "Cumsum:Bad 178\n",
      "Cumsum:Loan 38362700\n",
      "Cumsum:Mortdue 154048137.0\n",
      "Cumsum:Value 215722676.0\n",
      "Cumsum:Yoj 18170.0\n",
      "Cumsum:Derog 303.0\n",
      "Cumsum:Delinq 533.0\n",
      "Cumsum:Clage 364541.6500000003\n",
      "Cumsum:Ninq 2094.0\n",
      "Cumsum:Clno 44931.0\n",
      "Cumsum:Debtinc 68825.56199999999\n"
     ]
    }
   ],
   "source": [
    "print(\"CumSum\")\n",
    "print('Cumsum:Bad',data.bad.cumsum()[3549])\n",
    "print('Cumsum:Loan',data.loan.cumsum()[3549])\n",
    "print('Cumsum:Mortdue',data.mortdue.cumsum()[3549])\n",
    "print('Cumsum:Value',data.value.cumsum()[3549])\n",
    "print('Cumsum:Yoj',data.yoj.cumsum()[3549])\n",
    "print('Cumsum:Derog',data.derog.cumsum()[3549])\n",
    "print('Cumsum:Delinq',data.delinq.cumsum()[3549])\n",
    "print('Cumsum:Clage',data.clage.cumsum()[3549])\n",
    "print('Cumsum:Ninq',data.ninq.cumsum()[3549])\n",
    "print('Cumsum:Clno',data.clno.cumsum()[3549])\n",
    "print('Cumsum:Debtinc',data.debtinc.cumsum()[3549])\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CumProd\n",
      "CumProd:Bad 0\n",
      "CumProd:Loan 0\n",
      "CumProd:Mortdue inf\n",
      "CumProd:Value inf\n",
      "CumProd:Yoj 0.0\n",
      "CumProd:Derog 0.0\n",
      "CumProd:Delinq 0.0\n",
      "CumProd:Clage inf\n",
      "CumProd:Ninq 0.0\n",
      "CumProd:Clno nan\n",
      "CumProd:Debtinc inf\n"
     ]
    }
   ],
   "source": [
    "print(\"CumProd\")\n",
    "print('CumProd:Bad',data.bad.cumprod()[3549])\n",
    "print('CumProd:Loan',data.loan.cumprod()[3549])\n",
    "print('CumProd:Mortdue',data.mortdue.cumprod()[3549])\n",
    "print('CumProd:Value',data.value.cumprod()[3549])\n",
    "print('CumProd:Yoj',data.yoj.cumprod()[3549])\n",
    "print('CumProd:Derog',data.derog.cumprod()[3549])\n",
    "print('CumProd:Delinq',data.delinq.cumprod()[3549])\n",
    "print('CumProd:Clage',data.clage.cumprod()[3549])\n",
    "print('CumProd:Ninq',data.ninq.cumprod()[3549])\n",
    "print('CumProd:Clno',data.clno.cumprod()[3549])\n",
    "print('CumProd:Debtinc',data.debtinc.cumprod()[3549])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Describe\n",
      "               bad          loan        mortdue          value          yoj  \\\n",
      "count  2018.000000   2018.000000    2018.000000    2018.000000  2018.000000   \n",
      "mean      0.088206  19010.257681   76337.035183  106899.244797     9.003964   \n",
      "std       0.283665  10768.607345   44476.368189   54026.707096     7.573012   \n",
      "min       0.000000   1700.000000    5076.000000   23031.000000     0.000000   \n",
      "25%       0.000000  12000.000000   49577.250000   71287.250000     3.000000   \n",
      "50%       0.000000  16900.000000   67595.000000   94514.000000     7.000000   \n",
      "75%       0.000000  23600.000000   92922.750000  121220.750000    13.000000   \n",
      "max       1.000000  89900.000000  399412.000000  512650.000000    41.000000   \n",
      "\n",
      "             derog       delinq        clage         ninq         clno  \\\n",
      "count  2018.000000  2018.000000  2018.000000  2018.000000  2018.000000   \n",
      "mean      0.150149     0.264123   180.645020     1.037661    22.265114   \n",
      "std       0.580421     0.789816    84.420116     1.578414     9.457238   \n",
      "min       0.000000     0.000000     0.490000     0.000000     0.000000   \n",
      "25%       0.000000     0.000000   118.205000     0.000000    16.000000   \n",
      "50%       0.000000     0.000000   176.670000     0.000000    21.000000   \n",
      "75%       0.000000     0.000000   229.692500     2.000000    27.000000   \n",
      "max      10.000000    10.000000  1168.230000    13.000000    64.000000   \n",
      "\n",
      "           debtinc  \n",
      "count  2018.000000  \n",
      "mean     34.105829  \n",
      "std       8.438411  \n",
      "min       4.030000  \n",
      "25%      29.295500  \n",
      "50%      34.949000  \n",
      "75%      38.952000  \n",
      "max     144.189000   \n",
      " sayısal olmayan featureların istatistiksel özeti \n",
      "          reason    job\n",
      "count      2018   2018\n",
      "unique        2      6\n",
      "top     DebtCon  Other\n",
      "freq       1428    774 \n",
      " \n",
      " Job cesitleri  \n",
      " Other     774\n",
      "ProfEx    539\n",
      "Office    348\n",
      "Mgr       269\n",
      "Self       58\n",
      "Sales      30\n",
      "Name: job, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(\"Describe\")          \n",
    "print(data.describe(),\n",
    "'\\n','sayısal olmayan featureların istatistiksel özeti',\n",
    "'\\n',data.describe(include=['O']),#sayısal olmayan featureların istatistiksel özeti \n",
    "'\\n','\\n','Job cesitleri ','\\n',data.job.value_counts())\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Job ile Müşteri İlişkisi\n",
      "               Mgr  Office  Other  ProfEx  Sales  Self\n",
      "Bad Customer   242     326    694     504     23    51\n",
      "Good Customer   27      22     80      35      7     7\n",
      "                    Mgr    Office     Other    ProfEx     Sales      Self\n",
      "Bad Customer   0.119921  0.161546  0.343905  0.249752  0.011397  0.025273\n",
      "Good Customer  0.013380  0.010902  0.039643  0.017344  0.003469  0.003469\n",
      "\n",
      " Reason ile Müşteri İlişkisi\n",
      "               HomeImp  DebtCon\n",
      "Bad Customer      1304      536\n",
      "Good Customer      124       54\n",
      "                HomeImp   DebtCon\n",
      "Bad Customer   0.646184  0.265610\n",
      "Good Customer  0.061447  0.026759\n"
     ]
    }
   ],
   "source": [
    "#Sayısal olmayan feauterların ,sonuc ile ilişkisini gözlemlemek için cross table uygularız        \n",
    "#sayısal olmayan  bir feature olan job'ın \"bad\" feature ile cross table'ı;\n",
    "print(\"Job ile Müşteri İlişkisi\")          \n",
    "bad_job=pd.crosstab(index=data[\"bad\"],columns=data[\"job\"],margins=False)#margini false girersen satır ve kolon toplamını vermez true girersek verir \n",
    "bad_job.columns = [\"Mgr\",\"Office\",\"Other\",\"ProfEx\",\"Sales\",\"Self\"]#bad ile kıyaslanacak olan jobin seviyelerini gireriz\n",
    "bad_job.index=[\"Bad Customer\",\"Good Customer\"]#tablonun daha rahat  okunmasi icin iyi müşteri ve kötü müşteri  etiketi verdik\n",
    "print(bad_job)\n",
    "print(bad_job/len(data[\"job\"]))\n",
    "#sayısal olmayan diger bir feature olan reason'ın \"bad\" feature ile cross table'ı;\n",
    "print('\\n',\"Reason ile Müşteri İlişkisi\")  \n",
    "bad_reason=pd.crosstab(index=data[\"bad\"],columns=data[\"reason\"]) #bad müşteriler ile resason kolonu seviyesinin ilikisi \n",
    "bad_reason.columns = [\"HomeImp\",\"DebtCon\"]\n",
    "bad_reason.index=[\"Bad Customer\",\"Good Customer\"]\n",
    "print(bad_reason)\n",
    "print(bad_reason/len(data[\"reason\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<function matplotlib.pyplot.show(*args, **kw)>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#KORELASYON,ilişki matrisi,parametrelerin birbirleriyle olan ilişkisini verir \n",
    "corr=data.corr()\n",
    "corr.style.background_gradient(cmap='coolwarm')\n",
    "plt.matshow(data.corr())    \n",
    "plt.show"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "range(0, 2018)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Histogram of Debtinc Feature ')"
      ]
     },
     "execution_count": 14,
     "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#HISTOGRAM\n",
    "#datalarımızın featurelerını cizdirip hangi feature e göre ayrıstırabilirim diye bakabiliriz\n",
    "#verilerin tüm featurelarına bakarak (dagılımlarına bakarak) dagılımlara göre ayrıstırmayı deneyebilirz    \n",
    "#dataki tüm sayısal featureları cizdiririz;\n",
    "columns=data.columns.drop(['bad','job','reason'])    #datamızın ,\"bad\" hariç olan kolonlarını getir\n",
    "#create x data\n",
    "x_data=range(0,data.shape[0])\n",
    "#create figure and axists\n",
    "fig,ax=plt.subplots()\n",
    "for column in columns: #12 degisken icin verileri getir ve ciz deriz\n",
    "    ax.plot(x_data,data[column])\n",
    "    ax.set_title('Numerical Features')    \n",
    "    ax.legend()#bu legend ile hangi rengin hangi çiçeğe ait oldugunu yazarız \n",
    "print(x_data)\n",
    "\n",
    "#histogram cidirmek\n",
    "#loan\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"loan\"]   \n",
    "plt.hist(x,bins=50,color=\"green\")\n",
    "plt.title(\"Histogram of Loan Feature \")\n",
    "\n",
    "#mortdue\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"mortdue\"]   \n",
    "plt.hist(x,bins=50,color=\"blue\")\n",
    "plt.title(\"Histogram of Mortdue Feature \")\n",
    "\n",
    "#value\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"value\"]   \n",
    "plt.hist(x,bins=50,color=\"red\")\n",
    "plt.title(\"Histogram of Value Feature \")\n",
    "\n",
    "#yoj\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"yoj\"]   \n",
    "plt.hist(x,bins=50,color=\"purple\")\n",
    "plt.title(\"Histogram of Yoj Feature \")\n",
    "    \n",
    "#derog\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"derog\"]   \n",
    "plt.hist(x,bins=50,color=\"yellow\")\n",
    "plt.title(\"Histogram of Derog Feature \")\n",
    "    \n",
    "#delinq\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"delinq\"]   \n",
    "plt.hist(x,bins=50,color=\"brown\")\n",
    "plt.title(\"Histogram of Delinq Feature \")\n",
    "        \n",
    "#clage\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"clage\"]   \n",
    "plt.hist(x,bins=50,color=\"c\")\n",
    "plt.title(\"Histogram of Clage Feature \")\n",
    "        \n",
    "#ninq\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"ninq\"]   \n",
    "plt.hist(x,bins=50,color=\"black\")\n",
    "plt.title(\"Histogram of Ninq Feature \")\n",
    "        \n",
    "#clno\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"clno\"]   \n",
    "plt.hist(x,bins=50,color=\"orange\")\n",
    "plt.title(\"Histogram of Clno Feature \")\n",
    "            \n",
    "#debtinc\n",
    "plt.figure(figsize=(10,7))\n",
    "x=data[\"debtinc\"]   \n",
    "plt.hist(x,bins=50,color=\"red\")\n",
    "plt.title(\"Histogram of Debtinc Feature \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1eb2ff59c18>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf4AAAILCAYAAAAJ2P4SAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3XuYXXV97/H3hwwkgJSLaCAJEqzUk5ha60nR2rRNTFvFS/HpkUJqBSWVghjpwecQND2ltcYC51RPRSsNnUi8jSLVSgsIFGfqSVu5WS/A6CFFIGNSESFoAkGS/M4fewV3hkkmzG0P+b1fz7Of2eu3fmut756dnc9ev7VmrZRSkCRJddiv0wVIkqSJY/BLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfilSSTJFUne1+k6Om1Pv4ckb0mydoy205fkD8ZoXWNWlzSeDH5pCEnuTfJYks1JHk5yTZJjOl1XuyQlyQs6XYekZxaDX9q915dSngUcDXwfuLTD9YybtPj/gVQBP+jSMEopW4GrgLk725IcmuTjSX6Q5L4kf7wzOJN8NMlVbX0vTnJTE64LkwwkeU+SB5uRhTftbttJ3pZkXZKHklydZEbT/pWmyzeaUYlThlh2SpK/bLbz3STvaEYJupr5fUlWJvkX4FHg+UlmNNt5qNnu29rWt8vw+87X0jZ9b5J3J7mrGSX5WJJpbfNfl+TrSTYl+dckL26b94tJvpbkx0k+Czy53O5/Nbk0ySNJvp1kcdN4cpLbB3V8V5K/38O6fjbJLc26vpjkiLZlP5fkP5t5X0nyorZ5z25+Vz9Kcgvws8PULE0KBr80jCQHAacAX21rvhQ4FHg+8OvAacBbm3nvAl7cHPP9VWApcHr56fWxjwKOBGYCpwOrkrxwiO2+EvgL4HdpjTrcB3wGoJTya023XyilPKuU8tkhSn8bcCLwEuClwBuG6PNm4EzgkGb9PcAAMAN4I/D+naG6l94EvIpWCP4c8MfNa3kpsBr4Q+DZwN8AVyeZmuQA4O+BTwBHAJ8D/tsw23kZcA+t3+OFwOebwL4aOC7JnLa+v9+se3dOA86g9Zq3AR9qm3cdcDzwXOBrwKfa5n0E2ErrvTmjeUiTXynFhw8fgx7AvcBmYBOtMNgA/HwzbwrwODC3rf8fAn1t0ycAD9EK0yVt7Qub9R3c1nYl8D+b51cA72uedwOXtPV7FvAEMLuZLsAL9vAavgz8Ydv0bzTLdDXTfcB72+YfA2wHDmlr+wvgisG1tb2WgUG/s7Papl8D/Efz/KPAnw+q7zu0vjT9WvP7Tdu8f23f1qDl3jJE/1uAN7dta2Xz/EXAw8DU3ayrD7iobXou8BNgyhB9D2t+f4c2/waeAP5L2/z3A2s7/W/Xh4/hHu7xS7v3hlLKYcBU4B3APyfZubd+AK1Q3+k+WnvwAJRSbqG1Rxpawd7u4VLKlkHLzhhi+zPat1FK2Qz8sH07w5gBrG+bXj9En/a2GcBDpZQfD6ptb7c3eH3tr+tY4F3NMP+mJJtofdGY0Ty+V0opg5bdk6H679zWGuD3koTWiMaVpZTHn0bN+wNHNodKLkryH0l+ROuLDbTe/+cAXUMsK016Br80jFLK9lLK52ntDS8AHqS1t3dsW7fnAd/bOZHkHFpfGDYA5w9a5eFJDh607IYhNr2hfRvNMs9u384wNgKz2qaH+quE9vDcAByR5JBBte3c3hbgoLZ5Rw2xvvZttL+u9bT2wg9rexxUSulp6pzZBHX7snsyVP8NAKWUr9Laa/9V4PfY8zD/UDU/Qes9/j3gJFojJYcCs5s+AX5Aa+Rm8LLSpGfwS8NoTso7CTgc6C+lbKe1F78yySFJjgXOAz7Z9P854H20ji2/GTg/yUsGrfbPkhzQnAPwOlrHtQf7NPDWJC9JMpXWUPLNpZR7m/nfp3WOwe5cCZybZGaSw4Dle3qdpZT1tIbY/yLJtObku6X89Lj214HXJDmiGfn4oyFWc06SWc3x9vcAO889uBw4K8nLmt/nwUle23zJ+DdaIfrOJF1JfofWoZI9eW7Tf/8kJwNzgGvb5n8c+DCwrZQy3N/W/36Suc25HO8Frmre40NoHdL5Ia0vPO9v+11tBz4P/GmSg5LMpXW+hjTpGfzS7v1Dks3Aj4CVtE7Qu7OZt4zWHvA9wFpaIb26OWP+k8DFpZRvlFLuphWAn2jCG+A/aR133kArVM8qpXx78MZLKTcB/xP4O1p7xT8LnNrW5U+BNc3Q+e8OUf/lwA3AN4F/pxWM22iNXOzOElp7thuALwAXllJubOZ9AvgGrSHvG/hpqLf7dDPvnubxvua13EbrZMMPN699Ha1j9ZRSfgL8TjP9MK0TKT+/hxoBbqZ10t2DtN6bN5ZSftg2/xPAPIbf29/Z9wpa78s04J1N+8dpDd9/D7iLXU/uhNbhn2c1y10BfGwvtiV1XHY9TCZpPCVZCHyylDJruL7jsO0TgctKKccO23lk678X+INSyj+Nx/qfZi0HAg8AL22+fElquMcv7aOSHJjkNc3w+Uxaf/b2hU7XNUHOBm419KWn6up0AZLGTYA/ozUk/xhwDfAnHa1oAjQjD2Ho6xZI1XOoX5KkijjUL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEv6UlJ3pJkbafrkDR+DH5Jkipi8EuSVBGDX6pUkmOSfD7JD5L8MMmHh+jzV0nWJ/lRktuT/GrbvAOTrEnycJL+JOcnGWibPyPJ3zXr/26Sd07Ua5O0ewa/VKEkU4B/BO4DZgMzgc8M0fVW4CXAEcCngc8lmdbMu7BZ9vnAbwK/37b+/YB/AL7RrHsx8EdJXjX2r0bS02HwS3U6AZgB/I9SypZSytZSylNO6iulfLKU8sNSyrZSyl8CU4EXNrN/F3h/KeXhUsoA8KG2RX8JeE4p5b2llJ+UUu4BLgdOHddXJWlYXZ0uQFJHHAPcV0rZtqdOSd4F/AGtLwkF+BngyGb2DGB9W/f258cCM5JsamubAvzfUdYtaZQMfqlO64HnJenaXfg3x/OX0xqmv7OUsiPJw0CaLhuBWcBdzfQxg9b/3VLK8eNSvaQRc6hfqtMttIL7oiQHJ5mW5FcG9TkE2Ab8AOhK8ie09vh3uhJ4d5LDk8wE3jFo/T9Ksrw5CXBKknlJfmn8XpKkvWHwSxUqpWwHXg+8ALgfGABOGdTteuA64P/ROglwK7sO57+3We67wD8BVwGPD1r/S5r5DwJ/Cxw6Li9I0l5LKaXTNUjaByQ5Gzi1lPLrna5F0u65xy9pRJIcneRXkuyX5IXAu4AvdLouSXvmyX2SRuoA4G+A44BNtK4D8NcdrUjSsBzqlySpIg71S5JUEYNfkqSKTOpj/EceeWSZPXt2p8vQGNmyZQsHH3xwp8uQ1MbP5b7j9ttvf7CU8pzh+k3q4J89eza33XZbp8vQGOnr62PhwoWdLkNSGz+X+44k9+1NP4f6JUmqiMEvSVJFDH5JkioybPAnWZ3kgSR3tLX9ryTfTvLNJF9IcljbvHcnWZfkO0le1db+6qZtXZILxv6lSJKk4ezNHv8VwKsHtd0IzCulvJjWDTzeDZBkLnAq8KJmmb9u7so1BfgIcCIwF1jS9JUkSRNo2OAvpXwFeGhQ2w1t9/D+Kq17cgOcBHymlPJ4KeW7wDrghOaxrpRyTynlJ7Qu7XnSGL0GSZK0l8biGP8ZtG7dCTCTXW/bOdC07a5dkiRNoFH9HX+SFcA24FM7m4boVhj6C8aQNwlIciZwJsD06dPp6+sbTYmaRDZv3uz7KU0yfi7rM+LgT3I68DpgcfnpnX4GgGPaus0CNjTPd9e+i1LKKmAVwPz584sXlth3eKEQafLxc1mfEQ31J3k1sBz47VLKo22zrgZOTTI1yXHA8cAtwK3A8UmOS3IArRMArx5d6ZIk6ekado8/SQ+wEDgyyQBwIa2z+KcCNyYB+Gop5axSyp1JrgTuonUI4JxSyvZmPe8ArgemAKtLKXeOw+uRJEl7MGzwl1KWDNHcvYf+K4GVQ7RfC1z7tKqTJEljyiv3SZJUEYNfkqSKGPySVKGenh7mzZvH4sWLmTdvHj09PZ0uSRNkVH/HL0l65unp6WHFihV0d3ezfft2pkyZwtKlSwFYsmSo07q0L3GPX5Iqs3LlSrq7u1m0aBFdXV0sWrSI7u5uVq58ynnZ2gcZ/JJUmf7+fhYsWLBL24IFC+jv7+9QRZpIBr8kVWbOnDmsXbt2l7a1a9cyZ86cDlWkiWTwS1JlVqxYwdKlS+nt7WXbtm309vaydOlSVqxY0enSNAE8uU+SKrPzBL5ly5bR39/PnDlzWLlypSf2VcLgl6QKLVmyhCVLlniTngo51C9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSarIsMGfZHWSB5Lc0dZ2RJIbk9zd/Dy8aU+SDyVZl+SbSV7atszpTf+7k5w+Pi9HkiTtyd7s8V8BvHpQ2wXATaWU44GbmmmAE4Hjm8eZwEeh9UUBuBB4GXACcOHOLwuSJGniDBv8pZSvAA8Naj4JWNM8XwO8oa3946Xlq8BhSY4GXgXcWEp5qJTyMHAjT/0yIUmSxtlIj/FPL6VsBGh+Prdpnwmsb+s30LTtrl2SJE2grjFeX4ZoK3tof+oKkjNpHSZg+vTp9PX1jVlx6qzNmzf7fkqTjJ/L+ow0+L+f5OhSysZmKP+Bpn0AOKat3yxgQ9O+cFB731ArLqWsAlYBzJ8/vyxcuHCobnoG6uvrw/dTmlz8XNZnpEP9VwM7z8w/HfhiW/tpzdn9LwceaQ4FXA/8VpLDm5P6fqtpkyRJE2jYPf4kPbT21o9MMkDr7PyLgCuTLAXuB05uul8LvAZYBzwKvBWglPJQkj8Hbm36vbeUMviEQUmSNM6GDf5SypLdzFo8RN8CnLOb9awGVj+t6iRJ0pjyyn2SJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUZVfAn+e9J7kxyR5KeJNOSHJfk5iR3J/lskgOavlOb6XXN/Nlj8QIkSdLeG3HwJ5kJvBOYX0qZB0wBTgUuBj5YSjkeeBhY2iyyFHi4lPIC4INNP0mSNIFGO9TfBRyYpAs4CNgIvBK4qpm/BnhD8/ykZppm/uIkGeX2JUnS09A10gVLKd9L8r+B+4HHgBuA24FNpZRtTbcBYGbzfCawvll2W5JHgGcDD7avN8mZwJkA06dPp6+vb6QlapLZvHmz76c0yfi5rM+Igz/J4bT24o8DNgGfA04comvZucge5v20oZRVwCqA+fPnl4ULF460RE0yfX19+H5Kk4ufy/qMZqj/N4DvllJ+UEp5Avg88ArgsGboH2AWsKF5PgAcA9DMPxR4aBTblyRJT9Nogv9+4OVJDmqO1S8G7gJ6gTc2fU4Hvtg8v7qZppn/5VLKU/b4JUnS+Blx8JdSbqZ1kt7XgG8161oFLAfOS7KO1jH87maRbuDZTft5wAWjqFuSJI3AiI/xA5RSLgQuHNR8D3DCEH23AiePZnuSJGl0vHKfJEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBr3HX09PDvHnzWLx4MfPmzaOnp6fTJUlStbo6XYD2bT09PaxYsYLu7m62b9/OlClTWLp0KQBLlizpcHWSVB/3+DWuVq5cSXd3N4sWLaKrq4tFixbR3d3NypUrO12aJFXJ4Ne46u/vZ2BgYJeh/oGBAfr7+ztdmiRVyaF+jasZM2awfPlyPvWpTz051P+mN72JGTNmdLo0SaqSwa9x9+ijj3LGGWdw//3387znPY9HH32UQw45pNNlSVKVHOrXuPre977Hjh07hvwpSZp47vFrXE2ZMoUtW7awY8cOAJ544oknh/wlSRPP4Ne42rZt21PaduzY8eQXAUnSxHKoXxPiqKOOYr/99uOoo47qdCmSVDX3+DXuDj74YD796U8/OcT/+te/ni1btnS6LEmqksGvcffYY4/xqle9iieeeIL999+f7du3d7okSaqWQ/0adzt27HjyZL4pU6Z4fF+SOsjg17g64ogjANi6desuP3e2S5ImlsGvcbVp0ybOPvtspk6dCsDUqVM5++yz2bRpU4crk6Q6GfwaV3PmzOHkk09m69at9Pb2snXrVk4++WTmzJnT6dIkqUqjCv4khyW5Ksm3k/Qn+eUkRyS5Mcndzc/Dm75J8qEk65J8M8lLx+YlaDJbsWIFS5cupbe3l23bttHb28vSpUtZsWJFp0uTpCqN9qz+vwK+VEp5Y5IDgIOA9wA3lVIuSnIBcAGwHDgROL55vAz4aPNT+7AlS5YAsGzZMvr7+5kzZw4rV658sl2SNLFGvMef5GeAXwO6AUopPymlbAJOAtY03dYAb2ienwR8vLR8FTgsydEjrlzPGEuWLOGOO+7gpptu4o477jD0JamDRjPU/3zgB8DHkvx7kr9NcjAwvZSyEaD5+dym/0xgfdvyA02bJEmaIKMZ6u8CXgosK6XcnOSvaA3r706GaCtP6ZScCZwJMH36dPr6+kZRoiaTzZs3+35Kk4yfy/qMJvgHgIFSys3N9FW0gv/7SY4upWxshvIfaOt/TNvys4ANg1daSlkFrAKYP39+Wbhw4ShK1GTS19eH76c0ufi5rM+Ih/pLKf8JrE/ywqZpMXAXcDVwetN2OvDF5vnVwGnN2f0vBx7ZeUhAkiRNjNGe1b8M+FRzRv89wFtpfZm4MslS4H7g5KbvtcBrgHXAo01fSZI0gUYV/KWUrwPzh5i1eIi+BThnNNuTJEmj45X7JEmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPwadz09PcybN4/Fixczb948enp6Ol2SJFXL4Ne46unp4dxzz2XLli0AbNmyhXPPPdfwl6QOMfg1rs4//3y6urpYvXo1119/PatXr6arq4vzzz+/06VJUpUMfo2rgYEB1qxZw6JFi+jq6mLRokWsWbOGgYGBTpcmSVUy+CVJqojBr3E1a9YsTjvtNHp7e9m2bRu9vb2cdtppzJo1q9OlSVKVujpdgPZtl1xyCeeeey5nnHEG9913H8ceeyzbt2/nAx/4QKdLk6QqucevcbVkyRJOOeUUNm7cSCmFjRs3csopp7BkyZJOlyZJVXKPX+Oqp6eHa665huuuu47t27czZcoUli5dyite8QrDX5I6wD1+jauVK1fS3d29y1n93d3drFy5stOlSVKVDH6Nq/7+fhYsWLBL24IFC+jv7+9QRZJUN4Nf42rOnDmsXbt2l7a1a9cyZ86cDlUkSXXzGL/G1YoVK3jta1/LY4899mTbgQceSHd3dwerkqR6jXqPP8mUJP+e5B+b6eOS3Jzk7iSfTXJA0z61mV7XzJ892m1r8rviiit2CX2Axx57jCuuuKIzBUlS5cZiqP9coP2A7cXAB0spxwMPA0ub9qXAw6WUFwAfbPppH3fDDTcAMGXKlF1+7myXJE2sUQV/klnAa4G/baYDvBK4qumyBnhD8/ykZppm/uKmvypwySWXcN1113HJJZd0uhRJqtpo9/j/D3A+sKOZfjawqZSyrZkeAGY2z2cC6wGa+Y80/bWPO+GEEzjvvPOYNm0a5513HieccEKnS5Kkao345L4krwMeKKXcnmThzuYhupa9mNe+3jOBMwGmT59OX1/fSEvUBFq0aNFu591yyy0MNbgzVFtvb++Y1iVpzzZv3uz/s5UZzVn9vwL8dpLXANOAn6E1AnBYkq5mr34WsKHpPwAcAwwk6QIOBR4avNJSyipgFcD8+fPLwoULR1GiJkopT/kOB7SO6e/YseMp7fvttx/bt28f77IkDaOvrw//n63LiIf6SynvLqXMKqXMBk4FvlxKeRPQC7yx6XY68MXm+dXNNM38L5fdpYX2GW9/+9tJQldX6ztmV1cXSXj729/e4cokqU7jcQGf5cB5SdbROoa/8w+2u4FnN+3nAReMw7Y1yVx66aWcc845u5zVf84553DppZd2uDJJqlMm8073/Pnzy2233dbpMjRGZl9wDfde9NpOlyGpjUP9+44kt5dS5g/Xz0v2SpJUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5JqtCyZcuYNm0aixYtYtq0aSxbtqzTJWmCjOZa/ZKkZ6Bly5Zx2WWXcfHFFzN37lzuuusuli9fDuBVNSvgHr8kVebyyy/n4osv3uV22RdffDGXX355p0vTBDD4Jakyjz/+OGedddYubWeddRaPP/54hyrSRDL4JakyU6dO5bLLLtul7bLLLmPq1KkdqkgTyWP8klSZt73tbU8e0587dy4f+MAHWL58+VNGAbRvMvglqTI7T+B7z3vew+OPP87UqVM566yzPLGvEg71S1KFLr30UrZu3Upvby9bt2419Cti8EuSVBGDX5KkiniMX5L2UUnGbF2llDFblzrLPX5J2keVUoZ9HLv8H/eqn/YdBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIoY/JIkVcTglySpIl2dLkCT3y/82Q088tgTY7Ku2RdcM6rlDz1wf75x4W+NSS2SVCODX8N65LEnuPei1456PX19fSxcuHBU6xjtFwdJqp1D/ZIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKjLi4E9yTJLeJP1J7kxybtN+RJIbk9zd/Dy8aU+SDyVZl+SbSV46Vi9CkiTtndHs8W8D3lVKmQO8HDgnyVzgAuCmUsrxwE3NNMCJwPHN40zgo6PYtiRJGoERB38pZWMp5WvN8x8D/cBM4CRgTdNtDfCG5vlJwMdLy1eBw5IcPeLKJUnS0zYmV+5LMhv4ReBmYHopZSO0vhwkeW7TbSawvm2xgaZt46B1nUlrRIDp06fT19c3FiVqlMbifdi8efOYrMd/E9LY8jNVl1EHf5JnAX8H/FEp5UdJdtt1iLbylIZSVgGrAObPn19Ge4lXjYEvXTPqS+3C2Fyyd6xqkfYFY3Ufjbd8acuolvceGs8sowr+JPvTCv1PlVI+3zR/P8nRzd7+0cADTfsAcEzb4rOADaPZviTVbCzuo+E9NOozmrP6A3QD/aWUD7TNuho4vXl+OvDFtvbTmrP7Xw48svOQgCRJmhij2eP/FeDNwLeSfL1pew9wEXBlkqXA/cDJzbxrgdcA64BHgbeOYtuSJGkERhz8pZS1DH3cHmDxEP0LcM5ItydJkkbPK/dJklSRMflzPu3bDplzAT+/5oLhO+6NNcN32XMtAKM7mUmSambwa1g/7r9o1GcOg2cPS9Jk4FC/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBHvzidJz1Bjdstsb5ddFYNfkp6hxuKW2d4uuz4O9UuSVBGDX5Kkihj8kiRVxGP82itjdgzvS6Nbz6EH7j82dUhSpQx+DWu0Jw/tNPuCa8ZsXZKkkXGoX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRfxzPkl6BhuTa2x4fY2qGPyS9Aw1FtfF8Poa9XGoX5Kkihj8kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvCIkRTLAAAENUlEQVRLklQRg1+SpIoY/JIkVcTglySpIga/JEkVMfglSaqIwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVxOCXJKkiEx78SV6d5DtJ1iW5YKK3L0lSzbomcmNJpgAfAX4TGABuTXJ1KeWuiaxDkmqQZO/6XTx8n1LKKKvRZDHRe/wnAOtKKfeUUn4CfAY4aYJrkKQqlFKGffT29u5VP+07JnSPH5gJrG+bHgBe1t4hyZnAmQDTp0+nr69vworTyC1atGiv+g23Z9Hb2zsG1UjaW5s3b/b/2cpMdPAPNe60y1fJUsoqYBXA/Pnzy8KFCyegLI3W3uwR9PX14fspTS5+Lusz0UP9A8AxbdOzgA0TXIMkSdWa6OC/FTg+yXFJDgBOBa6e4BokSarWhA71l1K2JXkHcD0wBVhdSrlzImuQJKlmE32Mn1LKtcC1E71dSZLklfskSaqKwS9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4JckqSIGvyRJFTH4JUmqiMEvSVJFDH5Jkipi8EuSVBGDX5Kkihj8kiRVJKWUTtewW0l+ANzX6To0Zo4EHux0EZJ24edy33FsKeU5w3Wa1MGvfUuS20op8ztdh6Sf8nNZH4f6JUmqiMEvSVJFDH5NpFWdLkDSU/i5rIzH+CVJqoh7/JIkVcTg17hL8uok30myLskFna5HEiRZneSBJHd0uhZNLINf4yrJFOAjwInAXGBJkrmdrUoScAXw6k4XoYln8Gu8nQCsK6XcU0r5CfAZ4KQO1yRVr5TyFeChTtehiWfwa7zNBNa3TQ80bZKkDjD4Nd4yRJt/SiJJHWLwa7wNAMe0Tc8CNnSoFkmqnsGv8XYrcHyS45IcAJwKXN3hmiSpWga/xlUpZRvwDuB6oB+4spRyZ2erkpSkB/g34IVJBpIs7XRNmhheuU+SpIq4xy9JUkUMfkmSKmLwS5JUEYNfkqSKGPySJFXE4Je0iySzR3rHttEsK2liGPySJFXE4Jc0lK4ka5J8M8lVSQ5K8idJbk1yR5JVSQKQ5L8m+UaSfwPO6XDdkoZh8EsayguBVaWUFwM/At4OfLiU8kullHnAgcDrmr4fA95ZSvnlzpQq6ekw+CUNZX0p5V+a558EFgCLktyc5FvAK4EXJTkUOKyU8s9N3090oFZJT0NXpwuQNCkNvpZ3Af4amF9KWZ/kT4FptG677HW/pWcQ9/glDeV5SXYO3S8B1jbPH0zyLOCNAKWUTcAjSRY08980sWVKerrc45c0lH7g9CR/A9wNfBQ4HPgWcC+t2y3v9FZgdZJHad2FUdIk5t35JEmqiEP9kiRVxOCXJKkiBr8kSRUx+CVJqojBL0lSRQx+SZIqYvBLklQRg1+SpIr8f/qF537hi1YEAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#BOXPLOT\n",
    "#bir parametrenin, iyi müşteri-kötü müşteri sonucuyla olan ilişkisini boxlplotla çizdircez ve \n",
    "#iyi-kötü müşteriliği etkileyen her degişkenlere böyle bakıcaz\n",
    "#parametrelerin,müşteri sonucu  için  birbirlerinden farkı varmı, o parametre onları farklılık anlamında etkiliyormu diye bakıcaz yani\n",
    "#birbirinden anlamlı düzeyde ayrıştırabilecegimiz parametreler varmı diye birbirleriyle ilişkisini incelicez.\n",
    "        \n",
    "data.boxplot(column=\"loan\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"mortdue\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"value\", by='bad', figsize=(8,8))\n",
    "#    data.boxplot(column=\"reason\", by='bad', figsize=(8,8))#sayısal olmayanların boxplotu cizilmez \n",
    "#    data.boxplot(column=\"job\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"yoj\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"derog\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"delinq\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"clage\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"ninq\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"clno\", by='bad', figsize=(8,8))\n",
    "data.boxplot(column=\"debtinc\", by='bad', figsize=(8,8))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loan 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "loan 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "mortdue 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "mortdue 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "value 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "value 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "yoj 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "yoj 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "derog 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "derog 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "delinq 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "delinq 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "clage 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "clage 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "ninq 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "ninq 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "clno 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "clno 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "debtinc 0  için  Dagılımı Normalden Gelmemektedir.h0 reject\n",
      "debtinc 1  için  Dagılımı Normalden Gelmemektedir.h0 reject\n"
     ]
    }
   ],
   "source": [
    "#SHAPIRO TEST\n",
    "#h0=Dagılım Normalden gelmektedir. h1=gelmemektedir.\n",
    "#her parametrenin bad kolonunun 0 degeri icin normal dagılıp dagılmadıgını \n",
    "#ve bad kolonunun 1 degerleri arasında normal dagılıp dagılmadıgına bakarız \n",
    "    \n",
    "        \n",
    "for key in my_dict:\n",
    "    for i in grps:\n",
    "            stat,p=shapiro(data[my_dict[key]][data['bad']==i])\n",
    "            alpha=0.05\n",
    "            if(p>alpha):\n",
    "                print(my_dict[key],i,\"için Dagılımı Normalden Gelmektedir.h0 accepted\")\n",
    "            else:\n",
    "                 print(my_dict[key],i,\" için  Dagılımı Normalden Gelmemektedir.h0 reject\")\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loan 0 için varyans homojendir .h0 accepted\n",
      "loan 1 için varyans homojendir .h0 accepted\n",
      "mortdue 0 için varyans homojendir .h0 accepted\n",
      "mortdue 1 için varyans homojendir .h0 accepted\n",
      "value 0 için varyans homojendir .h0 accepted\n",
      "value 1 için varyans homojendir .h0 accepted\n",
      "yoj 0 için varyans homojendir .h0 accepted\n",
      "yoj 1 için varyans homojendir .h0 accepted\n",
      "derog 0 için varyans homojendir .h0 accepted\n",
      "derog 1 için varyans homojendir .h0 accepted\n",
      "delinq 0 için varyans homojendir .h0 accepted\n",
      "delinq 1 için varyans homojendir .h0 accepted\n",
      "clage 0 için varyans homojendir .h0 accepted\n",
      "clage 1 için varyans homojendir .h0 accepted\n",
      "ninq 0 için varyans homojendir .h0 accepted\n",
      "ninq 1 için varyans homojendir .h0 accepted\n",
      "clno 0 için varyans homojendir .h0 accepted\n",
      "clno 1 için varyans homojendir .h0 accepted\n",
      "debtinc 0 için varyans homojendir .h0 accepted\n",
      "debtinc 1 için varyans homojendir .h0 accepted\n"
     ]
    }
   ],
   "source": [
    "#iki datanın varyansları farklımı diye bakıcaz constant(sabit,homojen) varyans vamı diye bakıyoruz\n",
    "       \n",
    "#barlett normal dagılımda daha iyidir.Leven normal olmayan  dagılımda daha iyidir.            \n",
    "#h0 varyanslar eşittir.\n",
    "#h1=varyanslardan en az bir tanesi farklıdır \n",
    "#biz h0 hipotezini isteriz\n",
    "for key in my_dict:\n",
    "    for i in grps:\n",
    "            stat,p=levene(data['loan'][data['bad']==0],data['loan'][data['bad']==1])\n",
    "            #bad kolonu 0 olanın loanı ile bad kolonu 1 olanın loanının varyanslarının farklı olupolmadıgına bakıyoruz p degerlerini alphayla kıyasalyarak.\n",
    "            alpha=0.05\n",
    "            if(p>alpha):\n",
    "                print(my_dict[key],i,\"için varyans homojendir .h0 accepted\")\n",
    "            else:\n",
    "                print(my_dict[key],i,\" için  varyans homojen degildir .h0 reject\")\n",
    "        \n",
    "           \n",
    "#DEBTİNC,MORTDUE,DELİNQ,CLAGE,DEROG,NİNQ,YOJ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loan 0 için anlamlı fark yoktur.h0 accepted\n",
      "loan 1 için anlamlı fark yoktur.h0 accepted\n",
      "mortdue 0 için anlamlı fark yoktur.h0 accepted\n",
      "mortdue 1 için anlamlı fark yoktur.h0 accepted\n",
      "value 0 için anlamlı fark yoktur.h0 accepted\n",
      "value 1 için anlamlı fark yoktur.h0 accepted\n",
      "yoj 0 için anlamlı fark yoktur.h0 accepted\n",
      "yoj 1 için anlamlı fark yoktur.h0 accepted\n",
      "derog 0 için anlamlı fark yoktur.h0 accepted\n",
      "derog 1 için anlamlı fark yoktur.h0 accepted\n",
      "delinq 0 için anlamlı fark yoktur.h0 accepted\n",
      "delinq 1 için anlamlı fark yoktur.h0 accepted\n",
      "clage 0 için anlamlı fark yoktur.h0 accepted\n",
      "clage 1 için anlamlı fark yoktur.h0 accepted\n",
      "ninq 0 için anlamlı fark yoktur.h0 accepted\n",
      "ninq 1 için anlamlı fark yoktur.h0 accepted\n",
      "clno 0 için anlamlı fark yoktur.h0 accepted\n",
      "clno 1 için anlamlı fark yoktur.h0 accepted\n",
      "debtinc 0 için anlamlı fark yoktur.h0 accepted\n",
      "debtinc 1 için anlamlı fark yoktur.h0 accepted\n"
     ]
    }
   ],
   "source": [
    "#T TEST\n",
    "for key in my_dict:\n",
    "    for i in grps:\n",
    "            stat,p=stats.ttest_ind(data['loan'][data['bad']==0],data['loan'][data['bad']==1],equal_var=True)\n",
    "            #bad kolonu 0 olanın loanı ile bad kolonu 1 olanın loanının varyanslarının farklı olupolmadıgına bakıyoruz p degerlerini alphayla kıyasalyarak.\n",
    "            alpha=0.05\n",
    "            if(p>alpha):\n",
    "                print(my_dict[key],i,\"için anlamlı fark yoktur.h0 accepted\")\n",
    "            else:\n",
    "                print(my_dict[key],i,\" için anlamlı fark vardır .h0 reject\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.001\n",
      "Model:                            OLS   Adj. R-squared:                  0.000\n",
      "Method:                 Least Squares   F-statistic:                     1.383\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):              0.240\n",
      "Time:                        12:52:24   Log-Likelihood:                -319.62\n",
      "No. Observations:                2018   AIC:                             643.2\n",
      "Df Residuals:                    2016   BIC:                             654.5\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.1013      0.013      7.908      0.000       0.076       0.126\n",
      "loan       -6.897e-07   5.86e-07     -1.176      0.240   -1.84e-06     4.6e-07\n",
      "==============================================================================\n",
      "Omnibus:                     1165.761   Durbin-Watson:                   0.008\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6302.214\n",
      "Skew:                           2.901   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.425   Cond. No.                     4.43e+04\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 4.43e+04. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.001\n",
      "Model:                            OLS   Adj. R-squared:                  0.000\n",
      "Method:                 Least Squares   F-statistic:                     1.271\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):              0.260\n",
      "Time:                        12:52:24   Log-Likelihood:                -319.68\n",
      "No. Observations:                2018   AIC:                             643.4\n",
      "Df Residuals:                    2016   BIC:                             654.6\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.1004      0.013      8.005      0.000       0.076       0.125\n",
      "mortdue    -1.601e-07   1.42e-07     -1.127      0.260   -4.39e-07    1.18e-07\n",
      "==============================================================================\n",
      "Omnibus:                     1165.975   Durbin-Watson:                   0.007\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6305.651\n",
      "Skew:                           2.902   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.428   Cond. No.                     1.76e+05\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 1.76e+05. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.001\n",
      "Model:                            OLS   Adj. R-squared:                  0.000\n",
      "Method:                 Least Squares   F-statistic:                     1.979\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):              0.160\n",
      "Time:                        12:52:24   Log-Likelihood:                -319.32\n",
      "No. Observations:                2018   AIC:                             642.6\n",
      "Df Residuals:                    2016   BIC:                             653.9\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.1058      0.014      7.557      0.000       0.078       0.133\n",
      "value      -1.644e-07   1.17e-07     -1.407      0.160   -3.94e-07    6.48e-08\n",
      "==============================================================================\n",
      "Omnibus:                     1165.523   Durbin-Watson:                   0.008\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6300.149\n",
      "Skew:                           2.900   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.425   Cond. No.                     2.66e+05\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 2.66e+05. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.003\n",
      "Model:                            OLS   Adj. R-squared:                  0.002\n",
      "Method:                 Least Squares   F-statistic:                     5.244\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):             0.0221\n",
      "Time:                        12:52:24   Log-Likelihood:                -317.69\n",
      "No. Observations:                2018   AIC:                             639.4\n",
      "Df Residuals:                    2016   BIC:                             650.6\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.1054      0.010     10.752      0.000       0.086       0.125\n",
      "yoj           -0.0019      0.001     -2.290      0.022      -0.004      -0.000\n",
      "==============================================================================\n",
      "Omnibus:                     1162.359   Durbin-Watson:                   0.011\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6256.073\n",
      "Skew:                           2.893   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.398   Cond. No.                         18.4\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.056\n",
      "Model:                            OLS   Adj. R-squared:                  0.055\n",
      "Method:                 Least Squares   F-statistic:                     118.6\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):           7.11e-27\n",
      "Time:                        12:52:24   Log-Likelihood:                -262.64\n",
      "No. Observations:                2018   AIC:                             529.3\n",
      "Df Residuals:                    2016   BIC:                             540.5\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.0709      0.006     11.184      0.000       0.058       0.083\n",
      "derog          0.1152      0.011     10.890      0.000       0.094       0.136\n",
      "==============================================================================\n",
      "Omnibus:                     1156.871   Durbin-Watson:                   0.117\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6445.075\n",
      "Skew:                           2.850   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.645   Cond. No.                         1.78\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.079\n",
      "Model:                            OLS   Adj. R-squared:                  0.079\n",
      "Method:                 Least Squares   F-statistic:                     172.8\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):           6.14e-38\n",
      "Time:                        12:52:24   Log-Likelihood:                -237.32\n",
      "No. Observations:                2018   AIC:                             478.6\n",
      "Df Residuals:                    2016   BIC:                             489.9\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.0615      0.006      9.630      0.000       0.049       0.074\n",
      "delinq         0.1009      0.008     13.147      0.000       0.086       0.116\n",
      "==============================================================================\n",
      "Omnibus:                     1143.097   Durbin-Watson:                   0.167\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6421.218\n",
      "Skew:                           2.798   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.711   Cond. No.                         1.46\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.008\n",
      "Model:                            OLS   Adj. R-squared:                  0.008\n",
      "Method:                 Least Squares   F-statistic:                     16.77\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):           4.38e-05\n",
      "Time:                        12:52:24   Log-Likelihood:                -311.96\n",
      "No. Observations:                2018   AIC:                             627.9\n",
      "Df Residuals:                    2016   BIC:                             639.1\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.1433      0.015      9.646      0.000       0.114       0.172\n",
      "clage         -0.0003   7.45e-05     -4.095      0.000      -0.000      -0.000\n",
      "==============================================================================\n",
      "Omnibus:                     1155.838   Durbin-Watson:                   0.022\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6183.176\n",
      "Skew:                           2.874   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.363   Cond. No.                         471.\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.012\n",
      "Model:                            OLS   Adj. R-squared:                  0.012\n",
      "Method:                 Least Squares   F-statistic:                     24.67\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):           7.38e-07\n",
      "Time:                        12:52:24   Log-Likelihood:                -308.04\n",
      "No. Observations:                2018   AIC:                             620.1\n",
      "Df Residuals:                    2016   BIC:                             631.3\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.0677      0.008      9.011      0.000       0.053       0.082\n",
      "ninq           0.0198      0.004      4.967      0.000       0.012       0.028\n",
      "==============================================================================\n",
      "Omnibus:                     1150.632   Durbin-Watson:                   0.028\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6129.824\n",
      "Skew:                           2.859   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.340   Cond. No.                         2.49\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.000\n",
      "Model:                            OLS   Adj. R-squared:                 -0.000\n",
      "Method:                 Least Squares   F-statistic:                    0.3202\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):              0.572\n",
      "Time:                        12:52:24   Log-Likelihood:                -320.15\n",
      "No. Observations:                2018   AIC:                             644.3\n",
      "Df Residuals:                    2016   BIC:                             655.5\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      0.0966      0.016      5.980      0.000       0.065       0.128\n",
      "clno          -0.0004      0.001     -0.566      0.572      -0.002       0.001\n",
      "==============================================================================\n",
      "Omnibus:                     1166.649   Durbin-Watson:                   0.006\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6314.182\n",
      "Skew:                           2.903   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.432   Cond. No.                         62.0\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                    bad   R-squared:                       0.063\n",
      "Model:                            OLS   Adj. R-squared:                  0.063\n",
      "Method:                 Least Squares   F-statistic:                     135.6\n",
      "Date:                Tue, 17 Sep 2019   Prob (F-statistic):           2.21e-30\n",
      "Time:                        12:52:24   Log-Likelihood:                -254.62\n",
      "No. Observations:                2018   AIC:                             513.2\n",
      "Df Residuals:                    2016   BIC:                             524.5\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept     -0.1996      0.025     -7.841      0.000      -0.250      -0.150\n",
      "debtinc        0.0084      0.001     11.646      0.000       0.007       0.010\n",
      "==============================================================================\n",
      "Omnibus:                     1137.642   Durbin-Watson:                   0.129\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             6126.860\n",
      "Skew:                           2.807   Prob(JB):                         0.00\n",
      "Kurtosis:                       9.430   Cond. No.                         146.\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "#OLS ,tek degişkenli regresyon gibidir bir bagımlı degişken ile bir bagımsız degişkeni acıklamaa calistigimiz durumdur bir çeşit anovadır\n",
    "#summarize fonksiyonunda sadece ona verilen iki kolonumu alıyoruz (x=loan y=bad gibi )\n",
    "#degişkenler arasında anlamlı fark varmı diye bakıyoruz   ordinary lee square \n",
    "       \n",
    "#aplha=0.05 ise;\n",
    "results=smf.ols('bad ~ (loan)',data).fit() \n",
    "print(results.summary())\n",
    "#p=0.240  h0 accepted yani loan parametresi anlamlı farklılık icermediğinden modelimize ilgili parametreyi katmayız  \n",
    "\n",
    "results=smf.ols('bad ~ mortdue',data).fit()      \n",
    "print(results.summary())\n",
    "#p=0.260  h0 accepted yani mortdue parametresi anlamlı farklılık icermediğinden modelimize ilgili parametreyi katmayız     \n",
    "\n",
    "results=smf.ols('bad ~ value',data).fit()    \n",
    "print(results.summary())    \n",
    "#p=0.160 h0 accepted yani value parametresi anlamlı farklılık icermediğinden modelimize ilgili parametreyi katmayız                \n",
    "            \n",
    "results=smf.ols('bad ~ yoj',data).fit()      \n",
    "print(results.summary())       \n",
    "#p=0.022 h1 accepted yani yoj parametresi anlamlı farklılık icerdiğinden modelimize ilgili parametreyi katarız      \n",
    "\n",
    "results=smf.ols('bad ~ derog',data).fit()      \n",
    "print(results.summary())   \n",
    "#p=0.000 h1 accepted yani derog parametresi anlamlı farklılık icerdiğinden modelimize ilgili parametreyi katarız\n",
    "\n",
    "results=smf.ols('bad ~ delinq',data).fit()       \n",
    "print(results.summary())   \n",
    "#p=0.000 h1 accepted yani delinq parametresi anlamlı farklılık icerdiğinden modelimize ilgili parametreyi katarız\n",
    "\n",
    "results=smf.ols('bad ~ clage',data).fit()       \n",
    "print(results.summary())   \n",
    "#p=0.000 h1 accepted yani clage parametresi anlamlı farklılık icerdiğinden modelimize ilgili parametreyi katarız\n",
    "\n",
    "results=smf.ols('bad ~ ninq',data).fit()       \n",
    "print(results.summary())   \n",
    "#p=0.000 h1 accepted yani ninq parametresi anlamlı farklılık icerdiğinden modelimize ilgili parametreyi katarız\n",
    "\n",
    "results=smf.ols('bad ~ clno',data).fit()        \n",
    "print(results.summary())   \n",
    "#p=0.572 h0 accepted yani clno parametresi anlamlı farklılık icermediğinden modelimize ilgili parametreyi katmayız  \n",
    "        \n",
    "results=smf.ols('bad ~ debtinc',data).fit() \n",
    "print(results.summary())   \n",
    "#p=0.000 h1 accepted yani debtinc parametresi anlamlı farklılık icerdiğinden modelimize ilgili parametreyi katarız"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}