from sklearn.datasets import load_iris  #導(dǎo)入鳶尾花數(shù)據(jù)庫
#下面語句導(dǎo)入留一法LeaveOneOut，交叉驗證評價指標(biāo)cross_val_score評價指標(biāo)
from sklearn.model_selection import LeaveOneOut, cross_val_score
#下面語句導(dǎo)入邏輯斯特回歸LogisticRegression
from sklearn.linear_model import LogisticRegression
iris = load_iris()	#將鳶尾花數(shù)據(jù)庫中的數(shù)據(jù)放入iris變量中
print('Iris labels:\n{}'.format(iris.target))	#輸出鳶尾花的目標(biāo)分類，共有三種分別表示為0，1，2
logreg = LogisticRegression()	#邏輯斯特回歸對象logreg
loout = LeaveOneOut()	#留一法對象loout
#下面的語句計算評價指標(biāo)并輸出到scores對象，采用cress_val_score方法，
#方法參數(shù)含義為，采用邏輯斯特回歸方法logreg擬合數(shù)據(jù)：iris.data，iris.target,
#交叉驗證的方法為留一法loout
scores = cross_val_score(logreg,iris.data,iris.target,cv=loout)
#下面語句輸出每個樣本的擬合評價結(jié)果
print("leave-one-out cross validation scores:{}".format(scores))
#下面語句輸出整體評價結(jié)果,即所有樣本的評價結(jié)果平均值
print("Mean score of leave-one-out cross validation:{:.2f}".format(scores.mean()))

#導(dǎo)入訓(xùn)練集數(shù)據(jù)劃分方法庫train_test_split、留p法庫LevePOut
from sklearn.model_selection import train_test_split, LeavePOut
from sklearn import datasets	#導(dǎo)入數(shù)據(jù)集
from sklearn import svm			#導(dǎo)入支持向量機(jī)svm
from sklearn.metrics import accuracy_score	#導(dǎo)入評價指標(biāo)庫
import numpy as np	#導(dǎo)入數(shù)學(xué)分析庫
iris = datasets.load_iris()		#將鳶尾花數(shù)數(shù)據(jù)放入iris變量中
clf_svc = svm.SVC(kernel='linear')	#線性核函數(shù)svm模型對象為clf_svc
lpo = LeavePOut(p=3)				#留p法對象lpo
lpo.get_n_splits(iris.data)		#對鳶尾花數(shù)據(jù)進(jìn)行劃分
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))

#以下幾行代碼，例行公事，載入需要的庫函數(shù)
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自動對乳腺癌數(shù)據(jù)進(jìn)行k折交叉驗證，
#注意：cross_val_score的參數(shù)cv沒有設(shè)置，默認(rèn)是3，可以修改cv為5或10，
#則該方法則成為5或10折交叉驗證
scores = cross_val_score(logreg,cancer.data,cancer.target)
#打印分析結(jié)果,由于默認(rèn)cv為3，即3折交叉驗證，所以第一個輸出為三個指標(biāo)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()))