使用 rstatix 识别多个变量异常值

Question

这是我所拥有的数据的dput。我只包含了数据的头部，因为这是一个相当大的数据集，但我认为考虑到我的问题，它应该足够了：

structure(list(Prioritising.workload = c(2L, 2L, 2L, 4L, 1L, 
2L), Writing.notes = c(5L, 4L, 5L, 4L, 2L, 3L), Workaholism = c(4L, 
5L, 3L, 5L, 3L, 3L), Reliability = c(4L, 4L, 4L, 3L, 5L, 3L), 
    Self.criticism = c(1L, 4L, 4L, 5L, 5L, 4L), Loneliness = c(3L, 
    2L, 5L, 5L, 3L, 2L), Changing.the.past = c(1L, 4L, 5L, 5L, 
    4L, 3L), Number.of.friends = c(3L, 3L, 3L, 1L, 3L, 3L), Mood.swings = c(3L, 
    4L, 4L, 5L, 2L, 3L), Socializing = c(3L, 4L, 5L, 1L, 3L, 
    4L), Energy.levels = c(5L, 3L, 4L, 2L, 5L, 4L), Interests.or.hobbies = c(3L, 
    3L, 5L, NA, 3L, 5L)), row.names = c(NA, 6L), class = "data.frame")

我正在尝试找到所有这些变量的异常值。如果我单独执行此操作，我将得到以下与尼罗河一样长的代码：

#### EFA Personality Data Check ####
ef.personality %>% 
  identify_outliers(Prioritising.workload) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Writing.notes) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Workaholism) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Reliability) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Self.criticism) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Loneliness) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Changing.the.past) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Number.of.friends) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Mood.swings) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Socializing) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Energy.levels) %>% 
  select(is.extreme)
ef.personality %>% 
  identify_outliers(Interests.or.hobbies) %>% 
  select(is.extreme)

是否有一些命令可以使这变得更简单？我正在考虑某种循环，可以检查每个变量并返回每个变量的异常值，但我不确定如何实现这一点。我也对不依赖 rstatix 的解决方案持开放态度。

Answer 1

rstatix 的优点在于它对管道友好。因此，您可以将它与 tidyverse 框架一起使用。 tidyverse 需要长格式的数据。您可以使用以下代码

library(tidyverse)
library(rstatix)

ef.personality %>% 
  mutate(id = seq(1, nrow(ef.personality),1)) %>% #To create a unique column required to make that data in long form 
  pivot_longer(-id) %>% #To make the data in long form required for `tidyverse`
  group_by(name) %>% #Based on which column you want aggregate 
  identify_outliers(value) %>% 
  select(name, is.extreme)