如何将连续变量重新编码为范围

Question

我需要将连续变量重新编码为类别，通常我使用“cut”函数，但在 cut 函数中我需要指定中断。我正在寻找一种方法来根据我的数据框中的其他分类变量设置不同的中断。

我的示例中的变量是 Cost，“breaks”在第二个表“cost.range”中，每个“Region”和每个“Category”都有一组不同的 Breaks

例子：

Region    Product     Category Cost
Country A Product 1     CAT A  731
Country B Product 1     CAT A  659
Country C Product 1     CAT A  385
Country D Product 1     CAT A  763
Country A Product 2     CAT A  701
Country B Product 2     CAT A  759
Country C Product 2     CAT A  580
Country D Product 2     CAT A  147
Country A Product 3     CAT B  645
Country B Product 3     CAT B  657
Country C Product 3     CAT B  424


Region       Category  Cost.Range      Range
Country A      CAT A         10          R1
Country A      CAT A         50          R2
Country A      CAT A        200          R3
Country A      CAT A       1000          R4
Country A      CAT B         20          R1
Country A      CAT B        100          R2
Country A      CAT B        400          R3
Country A      CAT B       1500          R4

生成示例的代码：

Region <- c("Country A","Country B","Country C","Country D","Country A","Country B","Country C","Country D","Country A","Country B","Country C","Country D","Country A","Country B","Country C","Country D")
Product <- c("Product 1","Product 1","Product 1","Product 1","Product 2","Product 2","Product 2","Product 2","Product 3","Product 3","Product 3","Product 3","Product 4","Product 4","Product 4","Product 4")
Category <- c("CAT A","CAT A","CAT A","CAT A","CAT A","CAT A","CAT A","CAT A","CAT B","CAT B","CAT B","CAT B","CAT B","CAT B","CAT B","CAT B")
Cost <- c(731,659,385,763,701,759,580,147,645,657,424,34,850,463,160,550)

Table1 <- data.frame(Region, Product, Category, Cost)

Region <- c("Country A","Country A","Country A","Country A","Country A","Country A","Country A","Country A")
Category <- c("CAT A","CAT A","CAT A","CAT A","CAT B","CAT B","CAT B","CAT B")
Cost.range <- c(10,50,200,1000,20,100,400,1500)
Range <- c("R1","R1","R3","R4","R1","R2","R3","R4")

Table2 <- data.frame(Region, Category, Cost.range, Range)

Answer 1

这不是最优雅的解决方案（我有兴趣看到更好的方法），但它应该可以达到您正在寻找的结果。

dplyr 包中的 select() 和 distinct() 函数查找 Region 和 Category 的可能组合。这些组合用于对两个表进行子集化，并将cut() 函数应用于每个子集。

library('dplyr')
library('data.table')

dt1 <- data.table(Table1)
dt2 <- data.table(Table2)

t2d <- Table2 %>% select(Region, Category) %>% distinct

for(i in 1:nrow(t2d)){
  dt2_range_subset <- dt2[Region == as.character(t2d$Region[i]) 
                          & Category == t2d$Category[i], Cost.range]
  dt1[Region == as.character(t2d$Region[i]) & Category == t2d$Category[i],
      Cost_factor := cut(as.matrix(Cost), dt2_range_subset)]
}