Python实现Adaboost（decisiontree）

# -*- coding: utf-8 -*-
"""
Created on Thu Sep  7 17:17:22 2017

@author: piaodexin
"""

from sklearn import datasets
from sklearn.svm import LinearSVC
from sklearn import ensemble
from sklearn.model_selection import validation_curve
import matplotlib.pyplot as plt
import numpy as np

data=datasets.load_digits()
x=data.data
y=data.target

estimator_1=LinearSVC()
estimator_2=ensemble.AdaBoostClassifier(LinearSVC(),n_estimators=100,algorithm='SAMME')
estimator_2.get_params().keys()

validation_curve()
n=np.linspace(0.1,1,20)
train_score1,validation_score1=validation_curve(estimator_1,x,y,param_name='C',param_range=n,cv=3)
train_score2,validation_score2=validation_curve(estimator_2,x,y,param_name='base_estimator__C',param_range=n,cv=3)

n=np.linspace(0.1,1,20)

plt.grid()
plt.fill_between(n,train_score1.mean(1)-train_score1.std(1),
                 train_score1.mean(1)+train_score1.std(1),color='r',alpha=0.1)
plt.fill_between(n,validation_score1.mean(1)-validation_score1.std(1),
                 validation_score1.mean(1)+validation_score1.std(1),color='g',alpha=0.1)
plt.plot(n,train_score1.mean(1),c='r',label='train score')
plt.plot(n,validation_score1.mean(1),c='g',label='validation score')
plt.legend(loc='best')
plt.show()

plt.grid()
plt.fill_between(n,train_score2.mean(1)-train_score2.std(1),
                 train_score2.mean(1)+train_score2.std(1),color='r',alpha=0.1)
plt.fill_between(n,validation_score2.mean(1)-validation_score2.std(1),
                 validation_score2.mean(1)+validation_score2.std(1),color='g',alpha=0.1)
plt.plot(n,train_score2.mean(1),c='r',label='train score')
plt.plot(n,validation_score2.mean(1),c='g',label='validation score')
plt.legend(loc='best')
plt.show()