notebook.community

Edit and run

Shelter Animal Outcomes 11

Bagging





In [1]:



    


from sklearn.ensemble import BaggingClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB 
from sklearn import cross_validation
from sklearn.feature_selection import SelectKBest, chi2, RFECV
from sklearn.cross_validation import StratifiedKFold
from sklearn.grid_search import GridSearchCV
from sklearn.pipeline import make_pipeline
import pandas as pd



    











In [2]:



    


df_train = pd.read_csv('../Shelter_train.csv')
df_test = pd.read_csv('../Shelter_test.csv')



    











In [3]:



    


X = df_train.ix[:, :-1]
y = df_train.ix[:, -1]
df_test = df_test.drop('ID', 1)

Logistic Regression





In [4]:



    


clf = BaggingClassifier(LogisticRegression(), max_samples=0.5, max_features=0.5)
cross_validation.cross_val_score(clf, X, y, scoring="log_loss")



    


















    
Out[4]:








array([-1.03859505, -1.05642992, -1.08024019])

















In [5]:



    


X_new = SelectKBest(chi2, k=7).fit_transform(X, y)
cross_validation.cross_val_score(clf, X_new, y, scoring="log_loss")



    


















    
Out[5]:








array([-1.06820144, -1.07306632, -1.09494661])

















In [6]:



    


selector = SelectKBest(chi2, k=7)
predictor = make_pipeline(selector, clf)
predictor.fit(X, y)
predictions = predictor.predict_proba(df_test)
output = pd.DataFrame(predictions, columns=['Adoption', 'Died', 'Euthanasia', 'Return_to_owner', 'Transfer'])
output.index.names = ['ID']
output.index += 1
output.head()



    


















    
Out[6]:










  
    

      
      Adoption
      Died
      Euthanasia
      Return_to_owner
      Transfer
    

  
  
    

      1
      0.307454
      0.006654
      0.070184
      0.210140
      0.405568
    

    

      2
      0.468738
      0.001875
      0.044335
      0.319245
      0.165807
    

    

      3
      0.510772
      0.007821
      0.043575
      0.095783
      0.342049
    

    

      4
      0.339458
      0.007107
      0.063332
      0.189334
      0.400768
    

    

      5
      0.513114
      0.004399
      0.033684
      0.232800
      0.216002
    

  




















In [7]:



    


output.to_csv('../submission-Bagging-LogisticRegression.2.0.csv', index_label = 'ID')

SVC





In [8]:



    


svc = BaggingClassifier(SVC(), max_samples=0.5, max_features=0.5)
cross_validation.cross_val_score(svc, X, y, scoring="log_loss")



    


















    
Out[8]:








array([-9.96199163, -8.57309252, -8.83333762])

















In [9]:



    


X_new = SelectKBest(chi2, k=7).fit_transform(X, y)
cross_validation.cross_val_score(svc, X_new, y, scoring="log_loss")



    


















    
Out[9]:








array([-8.3853414 , -9.92669214, -8.38454693])

















In [10]:



    


selectorSVC = SelectKBest(chi2, k=7)
predictor = make_pipeline(selectorSVC, svc)
predictor.fit(X, y)
predictions = predictor.predict_proba(df_test)
output = pd.DataFrame(predictions, columns=['Adoption', 'Died', 'Euthanasia', 'Return_to_owner', 'Transfer'])
output.index.names = ['ID']
output.index += 1
output.head()



    


















    
Out[10]:










  
    

      
      Adoption
      Died
      Euthanasia
      Return_to_owner
      Transfer
    

  
  
    

      1
      0.4
      0.0
      0.0
      0.1
      0.5
    

    

      2
      1.0
      0.0
      0.0
      0.0
      0.0
    

    

      3
      0.7
      0.0
      0.0
      0.0
      0.3
    

    

      4
      0.3
      0.0
      0.0
      0.1
      0.6
    

    

      5
      0.9
      0.0
      0.0
      0.0
      0.1
    

  




















In [11]:



    


output.to_csv('../submission-Bagging-SVC.2.0.csv', index_label = 'ID')

Decision Tree





In [12]:



    


decisionTree = BaggingClassifier(DecisionTreeClassifier(), max_samples=0.5, max_features=0.5)
cross_validation.cross_val_score(decisionTree, X, y, scoring="log_loss")



    


















    
Out[12]:








array([-0.99228152, -1.04551678, -1.00705403])

















In [13]:



    


X_new = SelectKBest(chi2, k=4).fit_transform(X, y)
cross_validation.cross_val_score(decisionTree, X_new, y, scoring="log_loss")



    


















    
Out[13]:








array([-1.00795675, -0.99000558, -1.01431619])

















In [14]:



    


selectorDTree = SelectKBest(chi2, k=4)
predictor = make_pipeline(selectorDTree, decisionTree)
predictor.fit(X, y)
predictions = predictor.predict_proba(df_test)
output = pd.DataFrame(predictions, columns=['Adoption', 'Died', 'Euthanasia', 'Return_to_owner', 'Transfer'])
output.index.names = ['ID']
output.index += 1
output.head()



    


















    
Out[14]:










  
    

      
      Adoption
      Died
      Euthanasia
      Return_to_owner
      Transfer
    

  
  
    

      1
      0.252313
      0.005126
      0.077127
      0.233295
      0.432140
    

    

      2
      0.503346
      0.000730
      0.030821
      0.268996
      0.196107
    

    

      3
      0.504070
      0.005231
      0.054165
      0.147862
      0.288672
    

    

      4
      0.239439
      0.013327
      0.108950
      0.200081
      0.438203
    

    

      5
      0.497003
      0.002079
      0.055932
      0.221618
      0.223368
    

  




















In [15]:



    


output.to_csv('../submission-Bagging-DecisionTree.2.0.csv', index_label = 'ID')

KNN





In [18]:



    


knn = BaggingClassifier(KNeighborsClassifier(), max_samples=0.5, max_features=0.5)
cross_validation.cross_val_score(knn, X, y, scoring="log_loss")



    


















    
Out[18]:








array([-1.21654176, -1.34443792, -1.440332  ])

















In [19]:



    


X_new = SelectKBest(chi2, k=8).fit_transform(X, y)
cross_validation.cross_val_score(knn, X_new, y, scoring="log_loss")



    


















    
Out[19]:








array([-1.3564969 , -1.24266913, -1.34157001])

















In [20]:



    


selectorKNN = SelectKBest(chi2, k=8)
predictor = make_pipeline(selectorKNN, knn)
predictor.fit(X, y)
predictions = predictor.predict_proba(df_test)
output = pd.DataFrame(predictions, columns=['Adoption', 'Died', 'Euthanasia', 'Return_to_owner', 'Transfer'])
output.index.names = ['ID']
output.index += 1
output.head()



    


















    
Out[20]:










  
    

      
      Adoption
      Died
      Euthanasia
      Return_to_owner
      Transfer
    

  
  
    

      1
      0.26
      0.0
      0.02
      0.30
      0.42
    

    

      2
      0.56
      0.0
      0.10
      0.12
      0.22
    

    

      3
      0.56
      0.0
      0.04
      0.12
      0.28
    

    

      4
      0.36
      0.0
      0.16
      0.20
      0.28
    

    

      5
      0.50
      0.0
      0.04
      0.24
      0.22
    

  




















In [21]:



    


output.to_csv('../submission-Bagging-KNN.2.0.csv', index_label = 'ID')

Naive Bayes





In [22]:



    


gaussianNB = BaggingClassifier(GaussianNB(), max_samples=0.5, max_features=0.5)
cross_validation.cross_val_score(gaussianNB, X, y, scoring="log_loss")



    


















    
Out[22]:








array([-1.18349602, -1.20159273, -1.36292606])

















In [23]:



    


X_new = SelectKBest(chi2, k=4).fit_transform(X, y)
cross_validation.cross_val_score(gaussianNB, X_new, y, scoring="log_loss")



    


















    
Out[23]:








array([-1.02663551, -1.02429311, -1.02377623])

















In [24]:



    


selectorNB = SelectKBest(chi2, k=4)
predictor = make_pipeline(selectorNB, gaussianNB)
predictor.fit(X, y)
predictions = predictor.predict_proba(df_test)
output = pd.DataFrame(predictions, columns=['Adoption', 'Died', 'Euthanasia', 'Return_to_owner', 'Transfer'])
output.index.names = ['ID']
output.index += 1
output.head()



    


















    
Out[24]:










  
    

      
      Adoption
      Died
      Euthanasia
      Return_to_owner
      Transfer
    

  
  
    

      1
      0.141223
      0.004971
      0.086521
      0.274160
      0.493125
    

    

      2
      0.471021
      0.002381
      0.036342
      0.313238
      0.177017
    

    

      3
      0.552440
      0.009027
      0.047226
      0.055115
      0.336192
    

    

      4
      0.213122
      0.007911
      0.078015
      0.306603
      0.394349
    

    

      5
      0.507467
      0.001545
      0.029447
      0.319208
      0.142332
    

  




















In [25]:



    


output.to_csv('../submission-Bagging-Naive-Bayes.2.0.csv', index_label = 'ID')

Content source: krispingal/shelterAnimalOutcomes

Similar notebooks:

notebook.community | gallery | about