Image(graph.create_png())

from sklearn.model_selection import train_test_split
X_train,X_test, y_train, y_test = train_test_split(iris.data, iris.target,test_size=0.33,random_state=10)  
print(X_train.shape)
print(X_test.shape)

from sklearn import preprocessing
X_train_scaled = preprocessing.scale(X_train)
#print(X_trained_scaled)

print(X_train_scaled.mean(axis=0))
print(X_train_scaled.std(axis=0))

scaler = preprocessing.StandardScaler().fit(X_train) 
X_train_scaled = scaler.transform(X_train) 
print(X_train_scaled.mean(axis=0))
print(X_train_scaled.std(axis=0)) 
X_test_scaled = scaler.transform(X_test) 
print(X_test_scaled.mean(axis=0))
print(X_test_scaled.std(axis=0))

pipeline.get_params()

GridSearchCV??

from sklearn.model_selection import GridSearchCV
hyperparameters = { 
    'clf__max_depth': (2, 3, 4,5),  
    'clf__min_samples_split': ( 2,3,5),  
    'clf__min_samples_leaf': (1, 2,3)
}  
clf = GridSearchCV(pipeline, hyperparameters, cv=10,n_jobs=1,verbose=1)
clf.fit(X_train, y_train)

clf.best_params_

print ('最佳效果：%0.3f' %clf.best_score_)  
print('最优参数')  
print(clf.best_params_)
predictions = clf.predict(X_test)  

print(classification_report(y_test, predictions)) 
predictions

Image(graph.create_png())

predictions = besttree.predict(X_test)  
print(classification_report(y_test, predictions)) 

data = pd.read_excel("default of credit card clients.xls", skiprows=[0])
data.head()

from sklearn import tree
from sklearn.cross_validation import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import classification_report
from sklearn.pipeline import Pipeline
from sklearn.grid_search import GridSearchCV
import pydotplus
%matplotlib inline
from IPython.display import Image
y = data['default payment next month']
X = data.drop(["ID",'default payment next month'], axis=1)
X_train, X_test, y_train, y_test = train_test_split(X, y, 
                                                  test_size=0.2, 
                                                    random_state=123, 
                                                    stratify=y)
pipeline = Pipeline([('clf',tree.DecisionTreeClassifier(criterion='gini'))]) 
hyperparameters = { 
    'clf__max_depth': (10, 15, 20),  
    'clf__min_samples_split': ( 2,10,20),  
    'clf__min_samples_leaf': (1,5,10),
    'clf__max_features': (5,10)
}  
clf = GridSearchCV(pipeline, hyperparameters, cv=10,n_jobs=1,verbose=1)
clf.fit(X_train, y_train)
besttree=clf.best_estimator_.get_params()["clf"]
besttree.fit(X_train, y_train)

dot_data = tree.export_graphviz(besttree, out_file="temptree.dot",
                                 feature_names=X_train.columns,
                                filled=True, rounded=True,
                                special_characters=True)

dot_file=r'.\temptree.dot'
graph = pydotplus.graph_from_dot_file(dot_file) 

graph.write_pdf("default.pdf")

predictions = besttree.predict(X_test)  
print(classification_report(y_test, predictions)) 
print(besttree.feature_importances_)

print('Best Score: {}'.format(clf.best_score_))

y_test[1:10],predictions_prob[:,1]

fig = plt.figure(figsize=(10, 5))
ax1 = fig.add_subplot(121,  ylabel="查全率",xlabel="查准率")
ax1.plot(precision,recall)
ax2 = fig.add_subplot(122,  ylabel="True positive rate",xlabel="False positive rate")
ax2.plot(fpr,tpr)

from sklearn.ensemble import RandomForestClassifier
parameter_gridsearch = {
                 'max_depth' : [5,7],  #depth of each decision tree
                 'n_estimators': [50,40],  #count of decision tree      
       
                 }
randomforest = RandomForestClassifier()  
gridsearch = GridSearchCV(randomforest,            
                               param_grid=parameter_gridsearch
                                )


gridsearch.fit(X_train,y_train)  
 
parameters = gridsearch.best_params_
print(parameters)
print('Best Score: {}'.format(gridsearch.best_score_))
predictions = gridsearch.predict(X_test)  
print(classification_report(y_test, predictions))  

plt.figure(figsize=(10, 5))
plt.subplot(121,  ylabel="查全率",xlabel="查准率",title="precision recall curve")
plt.plot(precision1,recall1,label="random forrest")
plt.plot(precision,recall,color="red",label="CART")
plt.legend(loc="best")
plt.subplot(122,  ylabel="True positive rate",xlabel="False positive rate",title="ROC curve")
plt.plot(fpr1,tpr1,label="random forrest")
plt.plot(fpr,tpr,color="red",label="CART")
plt.legend(loc="best")