Python中sklearn实现交叉验证示例分析

from sklearn.datasets import load_iris  #导入鸢尾花数据库
#下面语句导入留一法LeaveOneOut，交叉验证评价指标cross_val_score评价指标
from sklearn.model_selection import LeaveOneOut, cross_val_score
#下面语句导入逻辑斯特回归LogisticRegression
from sklearn.linear_model import LogisticRegression
iris = load_iris()	#将鸢尾花数据库中的数据放入iris变量中
print('Iris labels:\n{}'.format(iris.target))	#输出鸢尾花的目标分类，共有三种分别表示为0，1，2
logreg = LogisticRegression()	#逻辑斯特回归对象logreg
loout = LeaveOneOut()	#留一法对象loout
#下面的语句计算评价指标并输出到scores对象，采用cress_val_score方法，
#方法参数含义为，采用逻辑斯特回归方法logreg拟合数据：iris.data，iris.target,
#交叉验证的方法为留一法loout
scores = cross_val_score(logreg,iris.data,iris.target,cv=loout)
#下面语句输出每个样本的拟合评价结果
print("leave-one-out cross validation scores:{}".format(scores))
#下面语句输出整体评价结果,即所有样本的评价结果平均值
print("Mean score of leave-one-out cross validation:{:.2f}".format(scores.mean()))

#导入训练集数据划分方法库train_test_split、留p法库LevePOut
from sklearn.model_selection import train_test_split, LeavePOut
from sklearn import datasets	#导入数据集
from sklearn import svm			#导入支持向量机svm
from sklearn.metrics import accuracy_score	#导入评价指标库
import numpy as np	#导入数学分析库
iris = datasets.load_iris()		#将鸢尾花数数据放入iris变量中
clf_svc = svm.SVC(kernel='linear')	#线性核函数svm模型对象为clf_svc
lpo = LeavePOut(p=3)				#留p法对象lpo
lpo.get_n_splits(iris.data)		#对鸢尾花数据进行划分
mean_accuracy_score_list = []
for train_index, test_index in lpo.split(iris.data):
	clf_svc.fit(iris.data[train_index], iris.target[train_index])
	prediction = clf_svc.predict(iris.data[test_index])
	mean_accuracy_score_list.append(accuracy_score(iris.target[test_index],prediction))
print(np.average(mean_accuracy_score_list))

#以下几行代码，例行公事，载入需要的库函数
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import cross_val_score
from sklearn.linear_model import LogisticRegression
cancer = load_breast_cancer()
logreg = LogisticRegression()
#以下指令中的cross_val_score自动对乳腺癌数据进行k折交叉验证，
#注意：cross_val_score的参数cv没有设置，默认是3，可以修改cv为5或10，
#则该方法则成为5或10折交叉验证
scores = cross_val_score(logreg,cancer.data,cancer.target)
#打印分析结果,由于默认cv为3，即3折交叉验证，所以第一个输出为三个指标值
print("Cross validation scores:{}".format(scores))
print("Mean cross validation score:{:2f}".format(scores.mean()))

from sklearn.datasets import load_iris
from sklearn.model_selection import ShuffleSplit,cross_val_score
from sklearn.linear_model import LogisticRegression
iris = load_iris()
shufspl = ShuffleSplit(train_size=.5,test_size=.4,n_splits=8) #迭代8次；
logreg = LogisticRegression()
scores = cross_val_score(logreg,iris.data,iris.target,cv=shufspl)
print("shuffle split cross validation scores:\n{}".format(scores))
print("Mean score of shuffle split cross validation:{:.2f}".format(scores.mean()))

from sklearn.datasets import load_iris
from sklearn.model_selection import ShuffleSplit,cross_val_score
from sklearn.linear_model import LogisticRegression
iris = load_iris()
shufspl = ShuffleSplit(train_size=.5,test_size=.4,n_splits=8) #迭代8次；
logreg = LogisticRegression()
scores = cross_val_score(logreg,iris.data,iris.target,cv=shufspl)
print("shuffle split cross validation scores:\n{}".format(scores))
print("Mean score of shuffle split cross validation:{:.2f}".format(scores.mean()))