Question

我想知道是否有比我正在做的事情更简单的方法来创建这些变量？我正在尝试将车辆类型变量的值转换为变量本身。

##Set up boolean values##
norm.knnN$gearbox[norm.knnN$gearbox=="automatic"] = 1
norm.knnN$gearbox[norm.knnN$gearbox=="manual"] = 0
norm.knnN$gearbox = as.numeric(norm.knnN$gearbox)

norm.knnN$bus= ifelse(norm.knnN$vehicleType=="bus",1,0)
norm.knnN$cabrio= ifelse(norm.knnN$vehicleType=="cabrio",1,0)
norm.knnN$coupe= ifelse(norm.knnN$vehicleType=="coupe",1,0)
norm.knnN$limousine= ifelse(norm.knnN$vehicleType=="limousine",1,0)
norm.knnN$otherCar= ifelse(norm.knnN$vehicleType=="other",1,0)
norm.knnN$small_car= ifelse(norm.knnN$vehicleType=="small_car",1,0)
norm.knnN$station_wagon= ifelse(norm.knnN$vehicleType=="station_wagon",1,0)
norm.knnN$suv= ifelse(norm.knnN$vehicleType=="suv",1,0)
norm.knnN$vehicleType = NULL


norm.knnN$cng= ifelse(norm.knnN$fuelType=="cng",1,0)
norm.knnN$diesel= ifelse(norm.knnN$fuelType=="diesel",1,0)
norm.knnN$electric= ifelse(norm.knnN$fuelType=="electric",1,0)
norm.knnN$hybrid= ifelse(norm.knnN$fuelType=="hybrid",1,0)
norm.knnN$lpg= ifelse(norm.knnN$fuelType=="lpg",1,0)
norm.knnN$otherFuel= ifelse(norm.knnN$fuelType=="other",1,0)
norm.knnN$petrol= ifelse(norm.knnN$fuelType=="petrol",1,0)
norm.knnN$fuelType = NULL


norm.knnN$audi= ifelse(norm.knnN$brand=="audi",1,0)
norm.knnN$bmw= ifelse(norm.knnN$brand=="bmw",1,0)
norm.knnN$mercedes_benz= ifelse(norm.knnN$brand=="mercedes_benz",1,0)
norm.knnN$opel= ifelse(norm.knnN$brand=="opel",1,0)
norm.knnN$volkswagen= ifelse(norm.knnN$brand=="volkswagen",1,0)
norm.knnN$brand = NULL


norm.knnN$notRepairedDamage[norm.knnN$notRepairedDamage=="yes"] = 1
norm.knnN$notRepairedDamage[norm.knnN$notRepairedDamage=="no"] = 0
norm.knnN$notRepairedDamage = as.numeric(norm.knnN$notRepairedDamage)

Answer 1

cobalt程序包中有一个名为splitfactor()的辅助函数，该函数将因子分解为伪变量。您将运行以下命令：

norm.knnN <- cobalt::splitfactor(norm.knnN, 
                                c("gearbox", "vehicleType", 
                                 "fuelType", "brand", "notRepairedDamage"), 
                                drop.first = "if2")

设置drop.first = "if2"使得如果一个因子具有两个值（例如"yes"和"no"），则将删除其中一个虚拟变量，因为它对另一个。

Answer 2

最好使用dplyr的语法。有一些文献可以学习如何使用语法。（https://genomicsclass.github.io/book/pages/dplyr_tutorial.html）

在语法上，我通常使用case_when函数创建多个条件，而不是使用ifelse函数。

norm.knnN$gearbox[norm.knnN$gearbox=="automatic"] = 1
norm.knnN$gearbox[norm.knnN$gearbox=="manual"] = 0

它将更改为我提到的这种语法。

 norm.knnN %>%
    mutate(gearbox = case_when(
    gearbox == "automatic" ~ 1,
    gearbox == "manual" ~ 0,
    TRUE ~ NA))

有没有更好的方式编写此if-else语句？

2 个答案: