Limitations on Categorical Variable

categorical

#1

Hi All,

Just wanted to know. Is there any limitations on the number of values(categories) that can be used in a categorical variable for performing linear/logistic regressions?

Regards,
Karthikeyan P


#2

Not really, as long as you can get enough observations in each combination.

Regards,
Kunal


#3

Thanks a lot Kunal Sir.


#4

Hi @kunal Sir,

One of my dataset variables has close to 100 categories. This is leading me to “aliased coefficients issue” or “perfect multi-collinearity” when building a glm model. So do you think that this is because I dont have enough observations in each combinations ?

Regards,
Karthikeyan P