convert low frequent levels of a categorical variable in "others" in R

Question

I have a categorical variable that I want to convert into dummies for a classification task. the problem is that some of the levels appear just a few times, thus they create problems of perfect-multicollinearity when I split my sample in training set and testing set.

How can I get rid of these levels in a quick and elegant way? Here is a simple example of the my data:

label   var_x
 1        1
 0        2
 1        1
 0        3
 1        2
 0        4
 0        5
 1        5
 1        1
 ....

Let's say that I want to keep only the levels that appear more than 1 (or any other number) I want to recode as "0" those cases and obtain something like this:

label   var_x
 1        1
 0        2
 1        1
 0        0
 1        2
 0        0
 0        5
 1        5
 1        1
 ....

Thank you for your help

Answer 1

One dplyr option could be:

df %>%
 add_count(var_x) %>%
 mutate(var_x = as.numeric(n > 1)*var_x) %>%
 select(-n)

  label var_x
  <int> <dbl>
1     1     1
2     0     2
3     1     1
4     0     0
5     1     2
6     0     0
7     0     5
8     1     5
9     1     1

And the same idea with base R:

as.numeric(with(data.frame(table(df$var_x)), Freq[match(df$var, Var1)]) > 1)*df$var_x