What learners are well suited for count data?

wessel

What learners are well suited for count data?

I'm assuming a small number of samples in the data.
E.g. 30.

Typically I would use Poisson regression, but this is not included in Rapid Miner.

Best regards,

Wessel

wessel

Maybe someone can mention tips and tricks instead of some learner?

For example, a trick that is often used is applying a log transformation.
This transforms counts (which are always positive numbers) into ranging from negative infinity to positive infinity.
This makes sure your regression never produces predictions that are impossible.