I'm currently working on a topic modeling project and want to determine the number of latent topics in the dataset. I've recently discovered the UMass coherence score as a suitable proxy. The issue is that I don't know how to calculate it in Stata. Does anyone know an easy way to calculate it in Stata or another program that is easy to use? I've tried to use Mallet without success.

More Addam Reynolds's questions See All
Similar questions and discussions