RV-Dependent Observations from Irregularly-Shaped Dataset

Is the motif known or you want to also model it as a latent variable in the model?