Day 91 - 앙상블(Ensamble) 사용하기

https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_breast_cancer.html#sklearn.datasets.load_breast_cancer

load_breast_cancer

 

데이터 전처리

from sklearn.datasets import load_breast_cancer
data = load_breast_cancer()
data.target[[10, 50, 85]]
list(data.target_names)

data.data

data.feature_names

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

X=pd.DataFrame(data.data,columns=data.feature_names)
y=data.target

X.corr()

X.describe()

Decision tree 분류 모델

from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test=train_test_split(X,y)

from sklearn.tree import DecisionTreeClassifier

model=DecisionTreeClassifier()
model.fit(X_train,y_train)

model.score(X_test,y_test)

pd.DataFrame(model.feature_importances_,index=data.feature_names)

재현율, 정확도, 정밀도 구하기

from sklearn.metrics import accuracy_score,recall_score,precision_score,f1_score

print('정확도:',accuracy_score(y_test,model.predict(X_test)))
print('재현율:',recall_score(y_test,model.predict(X_test)))
print('정밀도:',precision_score(y_test,model.predict(X_test)))
print('f1 스코어:',f1_score(y_test,model.predict(X_test)))

voting

from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.neighbors import KNeighborsClassifier

from sklearn.ensemble import VotingClassifier

model1=DecisionTreeClassifier()
model2=LogisticRegression(max_iter=10000)
model3=KNeighborsClassifier()

voting=VotingClassifier(
    estimators=[('DT',model1),('LR',model2),('KN',model3)],
    voting='hard' #'soft'
)
voting.fit(X_train,y_train)

voting.score(X_test,y_test)

# voting.predict_proba(X_test)
for model in voting.estimators_:
    print(model.score(X_test,y_test))

bagging

from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import GridSearchCV

model=RandomForestClassifier()

option={
    'n_estimators':[100,200,300],
    'max_depth':[3,5,7],
    'min_samples_split':[2,3,5],
    'n_jobs':[-1],
}
gs_cv=GridSearchCV(model,option)
gs_cv.fit(X_train,y_train)

gs_cv.score(X_test,y_test)

gs_cv.best_params_

gs_cv.best_estimator_

data.feature_names[gs_cv.best_index_]

gs_cv.best_score_

top10=pd.Series(gs_cv.best_estimator_.feature_importances_,index=gs_cv.feature_names_in_).sort_values(ascending=False).head(10)
top10

plt.barh(width=top10,y=top10.index)

boosting

Ada Boost

from sklearn.ensemble import AdaBoostClassifier

model=AdaBoostClassifier(algorithm='SAMME')
model.fit(X_train,y_train)

model.score(X_test,y_test)

Gradient Boost

from sklearn.ensemble import GradientBoostingClassifier

model=GradientBoostingClassifier()
model.fit(X_train,y_train)

model.score(X_test,y_test)

XG Boost

from xgboost import XGBClassifier

model=XGBClassifier()
model.fit(X_train,y_train)

model.score(X_test,y_test)