notebook.community

Edit and run

Load LC_Data





In [53]:



    


lc_data = pd.DataFrame.from_csv('./lc_dataframe(cleaning).csv')
lc_data = lc_data.reset_index()
lc_data.tail()



    


















    
Out[53]:










  
    

      
      loan_amnt
      int_rate
      emp_title
      emp_length
      home_ownership
      annual_inc
      verification_status
      issue_d
      loan_status
      desc
      purpose
      dti
      delinq_2yrs
      inq_last_6mths
      pub_rec
      revol_bal
      revol_util
      total_acc
      initial_list_status
    

  
  
    

      268131
      31050
      21.99
      1
      10
      1
      875000.0
      1
      12
      1
      0
      3
      9.66
      1
      0
      0
      25770
      79.3
      13
      0
    

    

      268132
      10800
      7.89
      1
      8
      1
      92400.0
      1
      12
      1
      0
      2
      19.62
      1
      0
      0
      9760
      68.7
      36
      1
    

    

      268133
      9000
      9.17
      1
      1
      1
      80000.0
      1
      12
      1
      0
      3
      3.97
      1
      0
      0
      6320
      51.8
      17
      0
    

    

      268134
      14400
      25.99
      0
      11
      4
      62000.0
      1
      12
      1
      0
      3
      16.88
      0
      1
      1
      5677
      45.1
      30
      0
    

    

      268135
      8000
      12.59
      1
      4
      3
      45000.0
      1
      12
      1
      0
      3
      26.21
      0
      0
      0
      9097
      50.8
      47
      1

Added Scaling (loan_amnt, annual_inc, revol_bal) on LC_Data





In [54]:



    


from sklearn.preprocessing import scale
x_loan_amnt = lc_data['loan_amnt'].reshape(-1,1)

scaled_loan_amnt = pd.DataFrame(np.hstack([scale(x_loan_amnt)]),
                 columns = ["scale(x_loan_amnt)"]
                 )

x_annual_inc = lc_data['annual_inc'].reshape(-1,1)
scaled_annual_inc = pd.DataFrame(np.hstack([scale(x_annual_inc)]),
                 columns = ["scale(x_annual_inc)"]
                 )

x_revol_bal = lc_data['revol_bal'].reshape(-1,1)
scaled_revol_bal = pd.DataFrame(np.hstack([scale(x_revol_bal)]),
                 columns = ["scale(x_revol_bal)"]
                 )

"""print(scaled_loan_amnt['scale(x_loan_amnt)'])
print(scaled_annual_inc['scale(x_annual_inc)'])
print(scaled_revol_bal['scale(x_revol_bal)'])"""



    


















    






/home/dockeruser/anaconda2/lib/python2.7/site-packages/ipykernel/__main__.py:2: FutureWarning:

reshape is deprecated and will raise in a subsequent release. Please use .values.reshape(...) instead

/home/dockeruser/anaconda2/lib/python2.7/site-packages/ipykernel/__main__.py:8: FutureWarning:

reshape is deprecated and will raise in a subsequent release. Please use .values.reshape(...) instead

/home/dockeruser/anaconda2/lib/python2.7/site-packages/ipykernel/__main__.py:13: FutureWarning:

reshape is deprecated and will raise in a subsequent release. Please use .values.reshape(...) instead











    
Out[54]:








"print(scaled_loan_amnt['scale(x_loan_amnt)'])\nprint(scaled_annual_inc['scale(x_annual_inc)'])\nprint(scaled_revol_bal['scale(x_revol_bal)'])"

New data setting (drop loan_amnt, annual_inc, revol_bal) + (shift loan_status dataset)





In [55]:



    


lc_data = lc_data.drop('loan_amnt', 1)
lc_data = pd.concat([lc_data, scaled_loan_amnt], axis = 1)

lc_data = lc_data.drop('annual_inc', 1)
lc_data = pd.concat([lc_data, scaled_annual_inc], axis = 1)

lc_data = lc_data.drop('revol_bal', 1)
lc_data = pd.concat([lc_data, scaled_revol_bal], axis = 1)

y = lc_data['loan_status']
lc_data = lc_data.drop('loan_status', 1)
lc_data = pd.concat([lc_data, y], axis = 1)

lc_data.tail()



    


















    
Out[55]:










  
    

      
      int_rate
      emp_title
      emp_length
      home_ownership
      verification_status
      issue_d
      desc
      purpose
      dti
      delinq_2yrs
      inq_last_6mths
      pub_rec
      revol_util
      total_acc
      initial_list_status
      scale(x_loan_amnt)
      scale(x_annual_inc)
      scale(x_revol_bal)
      loan_status
    

  
  
    

      268131
      21.99
      1
      10
      1
      1
      12
      0
      3
      9.66
      1
      0
      0
      79.3
      13
      0
      2.127000
      13.782613
      0.561689
      1
    

    

      268132
      7.89
      1
      8
      1
      1
      12
      0
      2
      19.62
      1
      0
      0
      68.7
      36
      1
      -0.352019
      0.343283
      -0.287872
      1
    

    

      268133
      9.17
      1
      1
      1
      1
      12
      0
      3
      3.97
      1
      0
      0
      51.8
      17
      0
      -0.572376
      0.130342
      -0.470413
      1
    

    

      268134
      25.99
      0
      11
      4
      1
      12
      0
      3
      16.88
      0
      1
      1
      45.1
      30
      0
      0.088696
      -0.178766
      -0.504533
      1
    

    

      268135
      12.59
      1
      4
      3
      1
      12
      0
      3
      26.21
      0
      0
      0
      50.8
      47
      1
      -0.694797
      -0.470701
      -0.323053
      1

Decision Tree data setting





In [56]:



    


feature_names = ["int_rate",	"emp_title", "emp_length",	"home_ownership",	"verification_status",	"issue_d",	"desc",	"purpose",	"dti",	"delinq_2yrs",	"inq_last_6mths",	"pub_rec",	"revol_util",	"total_acc",	"initial_list_status",	"scale(x_loan_amnt)",	"scale(x_annual_inc)",	"scale(x_revol_bal)"]
dfX = lc_data[feature_names]
dfy = lc_data['loan_status']
dfX.tail()



    


















    
Out[56]:










  
    

      
      int_rate
      emp_title
      emp_length
      home_ownership
      verification_status
      issue_d
      desc
      purpose
      dti
      delinq_2yrs
      inq_last_6mths
      pub_rec
      revol_util
      total_acc
      initial_list_status
      scale(x_loan_amnt)
      scale(x_annual_inc)
      scale(x_revol_bal)
    

  
  
    

      268131
      21.99
      1
      10
      1
      1
      12
      0
      3
      9.66
      1
      0
      0
      79.3
      13
      0
      2.127000
      13.782613
      0.561689
    

    

      268132
      7.89
      1
      8
      1
      1
      12
      0
      2
      19.62
      1
      0
      0
      68.7
      36
      1
      -0.352019
      0.343283
      -0.287872
    

    

      268133
      9.17
      1
      1
      1
      1
      12
      0
      3
      3.97
      1
      0
      0
      51.8
      17
      0
      -0.572376
      0.130342
      -0.470413
    

    

      268134
      25.99
      0
      11
      4
      1
      12
      0
      3
      16.88
      0
      1
      1
      45.1
      30
      0
      0.088696
      -0.178766
      -0.504533
    

    

      268135
      12.59
      1
      4
      3
      1
      12
      0
      3
      26.21
      0
      0
      0
      50.8
      47
      1
      -0.694797
      -0.470701
      -0.323053

Model Setting





In [71]:



    


from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(dfX, dfy, test_size=0.25, random_state=1)



    











In [80]:



    


from sklearn.tree import DecisionTreeClassifier
model = DecisionTreeClassifier(criterion='entropy', max_depth=8, min_samples_leaf=2).fit(X_train, y_train)



    











In [75]:



    


from sklearn import tree
with open('lc.dot','w') as f:
    f = tree.export_graphviz(model, out_file=f)
import os
os.unlink('lc.dot')
## 저장소 만들기 시도



    











In [78]:



    


import pydotplus
dot_data = tree.export_graphviz(model, out_file=None)
graph = pydotplus.graph_from_dot_data(dot_data)
graph.write_pdf('lc.pdf')
## Pdf 파일로 뽑아내기 시도



    


















    
Out[78]:








True

Model Display





In [90]:



    


from IPython.display import Image
dot_data = tree.export_graphviz(model, out_file=f,
                               feature_names=["int_rate",	"emp_title", "emp_length",	"home_ownership",	"verification_status",	"issue_d",	"desc",	"purpose",	"dti",	"delinq_2yrs",	"inq_last_6mths",	"pub_rec",	"revol_util",	"total_acc",	"initial_list_status",	"scale(x_loan_amnt)",	"scale(x_annual_inc)",	"scale(x_revol_bal)"],
                               class_names = 'loan_status', 
                               filled=True, rounded = True, 
                               special_characters = True)
graph = pydotplus.graph_from_dot_data(dot_data)
Image(graph.create_jpeg())



    


















    






dot: graph is too large for cairo-renderer bitmaps. Scaling by 0.731325 to fit











    
Out[90]:
























In [93]:



    


from sklearn.metrics import confusion_matrix

confusion_matrix(y_train, model.predict(X_train))



    


















    
Out[93]:








array([[  3450,  41886],
       [  2449, 153317]])

















In [94]:



    


confusion_matrix(y_test, model.predict(X_test))



    


















    
Out[94]:








array([[ 1033, 14046],
       [  982, 50973]])

성능 비교하기





In [95]:



    


from sklearn.metrics import classification_report

print(classification_report(y_train, model.predict(X_train)))



    


















    






             precision    recall  f1-score   support

          0       0.58      0.08      0.13     45336
          1       0.79      0.98      0.87    155766

avg / total       0.74      0.78      0.71    201102


















In [96]:



    


print(classification_report(y_test, model.predict(X_test)))



    


















    






             precision    recall  f1-score   support

          0       0.51      0.07      0.12     15079
          1       0.78      0.98      0.87     51955

avg / total       0.72      0.78      0.70     67034

Content source: shinys825/lc_project

Similar notebooks:

notebook.community | gallery | about