Existence of clusters in model: how to fit the model?

I was wonder that if our model, resulted in some clusters ( for example for different subtypes of ADHD or depression), how we can fit our model? Do we need to cluster at first and then create model? or What?

Depends a bit on what you are trying to model. If the model identifies clusters, that’s good, but perhaps the next step is to identify what features in the data drive the clustering.

1 Like