0
$\begingroup$

I am new to data science, and I would like to know whether I can use n-1 categorical variables instead of the original n, without risking to lose information.

At example: I have three categorical variables A,B,C, and I write C in function of A and B, therefore I have reduced the number of variables from 3 to 2. Can I apply this reduction without loss the information of the original 3 variables?

Improve this question
$\endgroup$

2 Answers 2

Reset to default
1
$\begingroup$

Not only can you do this, but, in most parameterizations, it is required that you do so. Usually, one level is taken as the reference and the others are compared to it. You can also compare all levels to the overall mean, but I find that less intuitive.

Improve this answer
$\endgroup$
0
$\begingroup$

Have you seen the Dummy coding article on Wikipedia?

In dummy coding, the reference group is assigned a value of 0 for each code variable, the group of interest for comparison to the reference group is assigned a value of 1 for its specified code variable, while all other groups are assigned 0 for that particular code variable.[2]

Improve this answer
$\endgroup$
1
  • $\begingroup$ I have read this article, but I wanted to be completely sure that it's still true when I'm implementing this on ANN, as in my case. $\endgroup$
    – Simone
    Commented May 17, 2017 at 12:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.