Dimension Reduction techniques on nonnormal count data
- Does anybody have any idea whether there are any techniques out
there for this? Kind of like a factor analysis for nonnormal count
Essentially, I have one (count data) dependent variable, and some
independent count data variables that I believe all contribute to a
latent variable. I want to be able to test for whether the
dependent variable and latent variable are related by inverse-U
relationship. Any help out there?