ValueError: n_splits=10 cannot be greater than the number of members in each class

Question

我正在尝试运行以下代码：

from sklearn.model_selection import StratifiedKFold 
X = ["hey", "join now", "hello", "join today", "join us now", "not today", "join this trial", " hey hey", " no", "hola", "bye", "join today", "no","join join"]
y = ["n", "r", "n", "r", "r", "n", "n", "n", "n", "r", "n", "n", "n", "r"]

skf = StratifiedKFold(n_splits=10)

for train, test in skf.split(X,y):  
    print("%s %s" % (train,test))

但是我收到以下错误：

ValueError: n_splits=10 cannot be greater than the number of members in each class.

我看过这里 scikit-learn error: The least populated class in y has only 1 member，但我仍然不确定我的代码有什么问题。

我的列表的长度都是 14 print(len(X)) print(len(y))。

我的部分困惑是我不确定 members 在这种情况下的定义以及 class 的定义。

问题：如何修复错误？什么是会员？什么是 class？（在这种情况下）

Answer 1

分层是指保持每个折叠中每个 class 的比例。因此，如果您的原始数据集有 3 个 classes，比例分别为 60%、20% 和 20%，那么分层将尝试在每个折叠中保持该比例。

在你的情况下，

X = ["hey", "join now", "hello", "join today", "join us now", "not today",
     "join this trial", " hey hey", " no", "hola", "bye", "join today", 
     "no","join join"]
y = ["n", "r", "n", "r", "r", "n", "n", "n", "n", "y", "n", "n", "n", "y"]

您共有 14 个样本（成员），其分布为：

class    number of members         percentage
 'n'        9                        64
 'r'        3                        22
 'y'        2                        14

因此 StratifiedKFold 将尝试在每次折叠中保持该比例。现在您指定了 10 次折叠 (n_splits)。所以这意味着在一个折叠中，为了 class 'y' 保持比例，至少有 2 / 10 = 0.2 个成员。但是我们不能给少于 1 个成员（样本），所以这就是它在那里抛出错误的原因。

如果您设置的不是 n_splits=10，而是 n_splits=2，那么它会起作用，因为 'y' 的成员数将为 2 / 2 = 1。对于n_splits = 10 要正常工作，每个 classes 至少需要 10 个样本。

ValueError: n_splits=10 cannot be greater than the number of members in each class

ValueError: n_splits=10 cannot be greater than the number of members in each class

python

scikit-learn

cross-validation