r - 如何从插入符号包拆分数据的createDataPartition功能？

Question

从文档中:

For bootstrap samples, simple random sampling is used.

For other data splitting, the random sampling is done within the levels of ywhen y is a factor in an attempt to balance the class distributions withinthe splits.

For numeric y, the sample is split into groups sections based on percentilesand sampling is done within these subgroups.

For createDataPartition, the number of percentiles is set via the groupsargument.

我不明白为什么需要这种“平衡”的东西。我认为我表面上了解它，但是任何其他见解都将真正有帮助。

Answer 1

这意味着，如果您有一个具有10000行的数据集ds

set.seed(42)
ds <- data.frame(values = runif(10000))

具有2个“类”且分布不均(9000与1000)

ds$class <- c(rep(1, 9000), rep(2, 1000))
ds$class <- as.factor(ds$class)
table(ds$class)
#    1    2 
# 9000 1000 

您可以创建一个示例，该示例尝试维护 factor类的比率/“平衡”。

dpart <- createDataPartition(ds$class, p = 0.1, list = F)
dsDP <- ds[dpart, ]
table(dsDP$class)
#   1   2 
# 900 100