如何通过保留一对交叉验证（LeavePOut）来拆分数据以进行二进制分类？答案

【问题标题】：How to split data with leave one pair out cross validation (LeavePOut) for binary classification?如何通过保留一对交叉验证（LeavePOut）来拆分数据以进行二进制分类？
【发布时间】：2020-12-21 14:12:08
【问题描述】：

我想在二元分类问题上应用离开一对交叉验证 (LPOCV)。对于被选为保持/测试对的每一对样本，它应该是每个二元类中的一个样本。

我的代码是这样的：

from sklearn.model_selection import LeavePOut
import numpy as np
X = np.array([[1, 2], [3, 4], [5, 6], [7, 8],[9,10]])
y = np.array([0,1,1,0,0])
lpo = LeavePOut(2)
print(lpo.get_n_splits(X))

print(lpo)
LeavePOut(p=2)
for train_index, test_index in lpo.split(X):
  print("TRAIN:", train_index, "TEST:", test_index)
  X_train, X_test = X[train_index], X[test_index]
  y_train, y_test = y[train_index], y[test_index]

输出如下：

LeavePOut(p=2)
TRAIN: [2 3 4] TEST: [0 1]
TRAIN: [1 3 4] TEST: [0 2]
TRAIN: [1 2 4] TEST: [0 3]
TRAIN: [1 2 3] TEST: [0 4]
TRAIN: [0 3 4] TEST: [1 2]
TRAIN: [0 2 4] TEST: [1 3]
TRAIN: [0 2 3] TEST: [1 4]
TRAIN: [0 1 4] TEST: [2 3]
TRAIN: [0 1 3] TEST: [2 4]
TRAIN: [0 1 2] TEST: [3 4]

测试对[0 3]和[0 4]属于同一类0。他们有什么方法可以将 X 数据与包含 0 类和 1 类样本的测试对分开吗？

【问题讨论】：

标签： python-3.x scikit-learn cross-validation

【解决方案1】：

我认为您可以调整您的代码，以便省略测试集仅包含一个类（例如类 0）的索引的所有折叠：

from sklearn.model_selection import LeavePOut
import numpy as np
X = np.array([[1, 2], [3, 4], [5, 6], [7, 8],[9,10]])
y = np.array([0,1,1,0,0])
lpo = LeavePOut(2)
print(lpo.get_n_splits(X))

print(lpo)
LeavePOut(p=2)
for train_index, test_index in lpo.split(X):
    for x in range(0,len(test_index)):
      for z in range(1,len(test_index)):
        if(y[test_index[x]] != y[test_index[z]]):
          print("TRAIN:", train_index, "TEST:", test_index)
          X_train, X_test = X[train_index], X[test_index]
          y_train, y_test = y[train_index], y[test_index]

那么输出是：

LeavePOut(p=2)
TRAIN: [2 3 4] TEST: [0 1]
TRAIN: [1 3 4] TEST: [0 2]
TRAIN: [0 2 4] TEST: [1 3]
TRAIN: [0 2 3] TEST: [1 4]
TRAIN: [0 1 4] TEST: [2 3]
TRAIN: [0 1 3] TEST: [2 4]

删除的折叠是测试集仅代表一个类别的折叠，即 [0 3]、[0 4]、 >[1 2] 和 [3 4]

【讨论】：