根据数值区间创建分类变量

Question

我有这个数据框

data<-data.frame(class1=c("A","A","A","A","A","A","A","A","B","B","B","B","B","B","B","B"),
                 class2=c(1,2,3,4,5,6,7,8,1,2,3,4,5,6,7,8),
                        observations=c(444,475, 531,560,650,668,705,717,456,876,123,47,249,180,500,654))

并且需要基于“class2”的2个单位间隔创建一个新的分类变量“class3”。如果 class2 介于 1 和 2 之间，则“class3”为 1，依此类推。 “class2”是连续的。

我可以使用定义的间隔创建一个新表，然后加入。

intv<-data.frame(class2=c(1,2,3,4,5,6,7,8,1,2,3,4,5,6,7,8),
                 class3=c(1,1,2,2,3,3,4,4,1,1,2,2,3,3,4,4))

data.2<-left_join(data,intv,by = join_by(class2)) 



> data.2
   class1 class2 observations class3
1       A      1          444      1
2       A      1          444      1
3       A      2          475      1
4       A      2          475      1
5       A      3          531      2
6       A      3          531      2
7       A      4          560      2
8       A      4          560      2
9       A      5          650      3
10      A      5          650      3
11      A      6          668      3
12      A      6          668      3
13      A      7          705      4
14      A      7          705      4
15      A      8          717      4
16      A      8          717      4
17      B      1          456      1
18      B      1          456      1
19      B      2          876      1
20      B      2          876      1
21      B      3          123      2
22      B      3          123      2
23      B      4           47      2
24      B      4           47      2
25      B      5          249      3
26      B      5          249      3
27      B      6          180      3
28      B      6          180      3
29      B      7          500      4
30      B      7          500      4
31      B      8          654      4
32      B      8          654      4

但是真实的数据框有很多观察结果，所以需要很多时间。

是否有一个功能可以自动执行此操作，仅指示间隔大小？

Answer 1

尝试这样：

data$class3 <- cut(data$class2, breaks = seq(0, max(data$class2)+1, by = 2), labels = FALSE)

Answer 2

这是你想要的吗？

findClass = \(x) {
  n = max(x)
  if((n %% 2L) == 0L) n = n + 1L
  findInterval(x, seq(1L, n, 2L))
}

给予

> findClass(data$class2)
 [1] 1 1 2 2 3 3 4 4 1 1 2 2 3 3 4 4

Answer 3

对于包含的示例数据，除以 2 并向上舍入应该足够了：

data<-data.frame(class1=c("A","A","A","A","A","A","A","A","B","B","B","B","B","B","B","B"),
                 class2=c(1,2,3,4,5,6,7,8,1,2,3,4,5,6,7,8),
                 observations=c(444,475, 531,560,650,668,705,717,456,876,123,47,249,180,500,654))

data$class3 <- ceiling(data$class2 / 2)
# if you need it to be categorical / factor :
data$class3_fct <- as.factor(data$class3)

head(data, n = 10)
#>    class1 class2 observations class3 class3_fct
#> 1       A      1          444      1          1
#> 2       A      2          475      1          1
#> 3       A      3          531      2          2
#> 4       A      4          560      2          2
#> 5       A      5          650      3          3
#> 6       A      6          668      3          3
#> 7       A      7          705      4          4
#> 8       A      8          717      4          4
#> 9       B      1          456      1          1
#> 10      B      2          876      1          1
str(data)
#> 'data.frame':    16 obs. of  5 variables:
#>  $ class1      : chr  "A" "A" "A" "A" ...
#>  $ class2      : num  1 2 3 4 5 6 7 8 1 2 ...
#>  $ observations: num  444 475 531 560 650 668 705 717 456 876 ...
#>  $ class3      : num  1 1 2 2 3 3 4 4 1 1 ...
#>  $ class3_fct  : Factor w/ 4 levels "1","2","3","4": 1 1 2 2 3 3 4 4 1 1 ...

^{创建于 2024-01-19，使用 reprex v2.0.2}