How to properly enter categorical variables in logistic regression in Matlab

Hi there,

I am running a logistic regression using glmfit with 2 continuous variables and one categorical (4 levels). I am a bit confused about how to enter the categorical variable properly? Is there a trick?

Thanks, Vishi

Answer 1

[the cyclist]

The DUMMYVAR command will create an array that you can use for categorical data.

For example,

 x_cat = dummyvar([1;2;3])

outputs an array for which each column corresponds to a category. You'll want to read "help dummyvar," especially the paragraph that explains that you will likely need to delete one variable that corresponds to the regression intercept.

[Hat tip to Peter Perkins for explaining this to me a while back.]