r - 从 mlogit 对象中提取自变量列表

Question

我正在使用以下包:

library(mlogit)

我的数据准备如下

data(CollegeDistance, package="AER")
testdata <- CollegeDistance
testdata$Dist[testdata$distance<0.4] <- 1
testdata$Dist[testdata$distance<1 & testdata$distance>=0.4] <- 2
testdata$Dist[testdata$distance<2.5 & testdata$distance>=1] <- 3
testdata$Dist[testdata$distance>=2.5] <- 4

这是我的模型

testmodel <- mlogit(as.formula(Dist ~ 1|urban + unemp + tuition|1), testdata, shape='wide', choice='Dist')

现在当我运行下面的代码时，结果很奇怪

as.character(attr(testmodel$formula, 'rhs')[[2]])

> as.character(attr(testmodel$formula, 'rhs')[[2]])
[1] "+"             "urban + unemp" "tuition"    

我期望的是这样的:

chr [1:3] "urban" "unemp" "tuition"     

Answer 1

您可以使用 all.vars 代替

all.vars(testmodel$formula) # return all the variables 
## "Dist"    "urban"   "unemp"   "tuition"
all.vars(testmodel$formula)[-1] # to remove the dependent variable
[1] "urban"   "unemp"   "tuition"

还有你用过的功能

all.vars(attr(testmodel$formula, 'rhs')[[2]])
## [1] "urban"   "unemp"   "tuition"