K折交叉验证：sklearn.model_selection.KFold(n_splits=3, shuffle=False, random_state=None)
思路：将训练/测试数据集划分n_splits个互斥子集，每次用其中一个子集当作验证集，剩下的n_splits-1个作为训练集，进行n_splits次训练和测试，得到n_splits个结果
注意点：对于不能均等份的数据集，其前n_samples % n_splits子集拥有n_samples // n_splits + 1个样本，其余子集都只有n_samples // n_splits样本
参数说明：
n_splits：表示划分几等份
shuffle：在每次划分时，是否进行洗牌
①若为Falses时，其效果等同于random_state等于整数，每次划分的结果相同
②若为True时，每次划分的结果都不一样，表示经过洗牌，随机取样的
random_state：随机种子数
属性：
①get_n_splits(X=None, y=None, groups=None)：获取参数n_splits的值
扫描二维码关注公众号，回复：
11122072 查看本文章
②split(X, y=None, groups=None)：将数据集划分成训练集和测试集，返回索引生成器
通过一个不能均等划分的栗子，设置不同参数值，观察其结果
①设置shuffle=False，运行两次，发现两次结果相同
1 In [1]: from sklearn.model_selection import KFold 2 ...: import numpy as np 3 ...: X = np.arange(24).reshape(12,2) 4 ...: y = np.random.choice([1,2],12,p=[0.4,0.6]) 5 ...: kf = KFold(n_splits=5,shuffle=False) 6 ...: for train_index , test_index in kf.split(X): 7 ...: print('train_index:%s , test_index: %s ' %(train_index,test_index)) 8 ...: 9 ...: 10 train_index:[ 3 4 5 6 7 8 9 10 11] , test_index: [0 1 2] 11 train_index:[ 0 1 2 6 7 8 9 10 11] , test_index: [3 4 5] 12 train_index:[ 0 1 2 3 4 5 8 9 10 11] , test_index: [6 7] 13 train_index:[ 0 1 2 3 4 5 6 7 10 11] , test_index: [8 9] 14 train_index:[0 1 2 3 4 5 6 7 8 9] , test_index: [10 11] 15 16 In [2]: from sklearn.model_selection import KFold 17 ...: import numpy as np 18 ...: X = np.arange(24).reshape(12,2) 19 ...: y = np.random.choice([1,2],12,p=[0.4,0.6]) 20 ...: kf = KFold(n_splits=5,shuffle=False) 21 ...: for train_index , test_index in kf.split(X): 22 ...: print('train_index:%s , test_index: %s ' %(train_index,test_index)) 23 ...: 24 ...: 25 train_index:[ 3 4 5 6 7 8 9 10 11] , test_index: [0 1 2] 26 train_index:[ 0 1 2 6 7 8 9 10 11] , test_index: [3 4 5] 27 train_index:[ 0 1 2 3 4 5 8 9 10 11] , test_index: [6 7] 28 train_index:[ 0 1 2 3 4 5 6 7 10 11] , test_index: [8 9]
②设置shuffle=True时，运行两次，发现两次运行的结果不同
1 In [3]: from sklearn.model_selection import KFold 2 ...: import numpy as np 3 ...: X = np.arange(24).reshape(12,2) 4 ...: y = np.random.choice([1,2],12,p=[0.4,0.6]) 5 ...: kf = KFold(n_splits=5,shuffle=True) 6 ...: for train_index , test_index in kf.split(X): 7 ...: print('train_index:%s , test_index: %s ' %(train_index,test_index)) 8 ...: 9 ...: 10 train_index:[ 0 1 2 4 5 6 7 8 10] , test_index: [ 3 9 11] 11 train_index:[ 0 1 2 3 4 5 9 10 11] , test_index: [6 7 8] 12 train_index:[ 2 3 4 5 6 7 8 9 10 11] , test_index: [0 1] 13 train_index:[ 0 1 3 4 5 6 7 8 9 11] , test_index: [ 2 10] 14 train_index:[ 0 1 2 3 6 7 8 9 10 11] , test_index: [4 5] 15 16 In [4]: from sklearn.model_selection import KFold 17 ...: import numpy as np 18 ...: X = np.arange(24).reshape(12,2) 19 ...: y = np.random.choice([1,2],12,p=[0.4,0.6]) 20 ...: kf = KFold(n_splits=5,shuffle=True) 21 ...: for train_index , test_index in kf.split(X): 22 ...: print('train_index:%s , test_index: %s ' %(train_index,test_index)) 23 ...: 24 ...: 25 train_index:[ 0 1 2 3 4 5 7 8 11] , test_index: [ 6 9 10] 26 train_index:[ 2 3 4 5 6 8 9 10 11] , test_index: [0 1 7] 27 train_index:[ 0 1 3 5 6 7 8 9 10 11] , test_index: [2 4] 28 train_index:[ 0 1 2 3 4 6 7 9 10 11] , test_index: [5 8] 29 train_index:[ 0 1 2 4 5 6 7 8 9 10] , test_index: [ 3 11]
③设置shuffle=True和random_state=整数，发现每次运行的结果都相同
1 In [5]: from sklearn.model_selection import KFold 2 ...: import numpy as np 3 ...: X = np.arange(24).reshape(12,2) 4 ...: y = np.random.choice([1,2],12,p=[0.4,0.6]) 5 ...: kf = KFold(n_splits=5,shuffle=True,random_state=0) 6 ...: for train_index , test_index in kf.split(X): 7 ...: print('train_index:%s , test_index: %s ' %(train_index,test_index)) 8 ...: 9 ...: 10 train_index:[ 0 1 2 3 5 7 8 9 10] , test_index: [ 4 6 11] 11 train_index:[ 0 1 3 4 5 6 7 9 11] , test_index: [ 2 8 10] 12 train_index:[ 0 2 3 4 5 6 8 9 10 11] , test_index: [1 7] 13 train_index:[ 0 1 2 4 5 6 7 8 10 11] , test_index: [3 9] 14 train_index:[ 1 2 3 4 6 7 8 9 10 11] , test_index: [0 5] 15 16 In [6]: from sklearn.model_selection import KFold 17 ...: import numpy as np 18 ...: X = np.arange(24).reshape(12,2) 19 ...: y = np.random.choice([1,2],12,p=[0.4,0.6]) 20 ...: kf = KFold(n_splits=5,shuffle=True,random_state=0) 21 ...: for train_index , test_index in kf.split(X): 22 ...: print('train_index:%s , test_index: %s ' %(train_index,test_index)) 23 ...: 24 ...: 25 train_index:[ 0 1 2 3 5 7 8 9 10] , test_index: [ 4 6 11] 26 train_index:[ 0 1 3 4 5 6 7 9 11] , test_index: [ 2 8 10] 27 train_index:[ 0 2 3 4 5 6 8 9 10 11] , test_index: [1 7] 28 train_index:[ 0 1 2 4 5 6 7 8 10 11] , test_index: [3 9] 29 train_index:[ 1 2 3 4 6 7 8 9 10 11] , test_index: [0 5]
④n_splits属性值获取方式
1 In [8]: kf.split(X) 2 Out[8]: <generator object _BaseKFold.split at 0x00000000047FF990> 3 4 In [9]: kf.get_n_splits() 5 Out[9]: 5 6 7 In [10]: kf.n_splits 8 Out[10]: 5