如何更改R中几个虚拟变量的值？

Question

I have 167 dummy variables amongst other variables in my dataframe. For creating a newdata for prediction, I wish to change the value of the first dummy variable to 1 and replace the values of all other variables with zero. My dummy variables are called district_code2, district_code3, district_code4 and so on. So I want to fix district_code2 to take the value 1 and all the others to take the value 0.

I created these dummy variables using factor and added them to my data by using model.matrix as in

 dummies = data.frame(model.matrix(~district_code, data=data_wht_81_09))
 #to get rid of the intercept
 dummies1<-dummies[,-1]

I need to have the dummies in my data because after I run the regression I donot wish to take into account the coeffiecients on all the dummies in my prediction. I want to plot the fitted value with respect to one variable holding all others at their mean. For district dummies this implies adding a constant to all the fitted values. Hence I want to put the values of all other dummy variables to 0. May be there is a more efficient way to do this.Below I am showing a sample of the object dummies.

 dput(head(dummies1,4))
 structure(list(district_code2 = c(0, 0, 0, 0), district_code3 = c(0, 
 0, 0, 0), district_code4 = c(0, 0, 0, 0), district_code5 = c(0, 
 0, 0, 0), district_code6 = c(0, 0, 0, 0), district_code7 = c(0,0, 0, 0), 

I am only displaying the first 6 variables. How can I do this? Many thanks in advance.

Answer 1

There is rarely any need to manipulate dummy variables yourself (R does that behind the scenes when you use factors), but, if it is absolutely needed, you can simply identify columns whose name start with discrict_code , and change their value: other columns will be left as is.

d <- data.frame( 
  district_code2 = c(0, 0, 0, 0), 
  district_code3 = c(0, 0, 0, 0), 
  district_code4 = c(0, 0, 0, 0), 
  district_code5 = c(0, 0, 0, 0), 
  district_code6 = c(0, 0, 0, 0), 
  district_code7 = c(0,0, 0, 0), 
  x = 1:4
)
library(stringr)
d[,str_detect(names(d), "^district_code[0-9]+")] <- 0
d[,1] <- 1
d