如何用R中另一个的分位数创建一个变量？

问题描述

我正在尝试使用“dplyr”命令 mutate 创建一个变量，该变量必须指示另一个变量的分位数。

例如：

TypeTag[this.type]

到目前为止，我已经创建了一个函数来识别和返回分位数作为一个因子，并且它确实有效

# 1.  Fake data:
data <- data.frame(
  "id" = seq(1:20),"score" = round(rnorm(20,30,20)))

# 2. Creating varaible 'Quantile_5'
data <-data %>% 
  mutate(Quntile_5 = ????)

但是，例如，如果我想创建一个变量“Quantile_100”作为指示每个观察值从 1 到 100 的哪个位置的因子（在较大数据集的上下文中），这不是一个很好的解决方案。有没有更简单的方法来创建这些五分位数变量？

解决方法

我希望这就是您要找的：

library(dplyr)

data <- data.frame(
  "id" = seq(1:20),"score" = round(rnorm(20,30,20)))


data %>%
  mutate(quantile100 = findInterval(score,quantile(score,probs = seq(0,1,0.01)),rightmost.closed = TRUE)) %>%
  slice_head(n = 10)

   id score quantile100
1   1    59          95
2   2    47          90
3   3    83         100
4   4    33          53
5   5     7          11
6   6    26          43
7   7    16          16
8   8    18          27
9   9    33          53
10 10    47          90

我选择关闭最右边的 bin，以便最大类别不超过 100。 我们也可以用你自己的例子来验证它，这会导致相同的结果：

df %>%
  mutate(quantile100 = findInterval(score,0.2)),rightmost.closed = TRUE)) %>%
  slice_head(n = 10)


   id score quantile5
1   1    55         5
2   2    56         5
3   3    26         3
4   4    42         3
5   5    41         3
6   6    26         3
7   7    57         5
8   8    12         1
9   9    21         2
10 10    25         2

数据

structure(list(id = 1:20,score = c(55L,56L,26L,42L,41L,57L,12L,21L,25L,37L,18L,54L,47L,52L,-4L,53L,51L,-7L,-2L)),class = "data.frame",row.names = c("1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20"))

,

这里有两个带有 cut 的选项：

1.

library(dplyr)

data %>% mutate(quantile100 = cut(score,100,label = FALSE))

#This is similar to @Anoushiravan R `findInterval` function.
data %>% 
    mutate(quantile100 = cut(score,unique(quantile(score,seq(0,0.01))),labels = FALSE))