An in-development R package and a Bayesian hierarchical model jointly fitting multiple "local" wastewater data streams and "global" case count data to produce nowcasts and forecasts of both observations
We are currently using a constrained simplex for the weekday effect with a normal prior for the mean of each element. It isn't clear how stan implements this under the hood, and would be better to sample directly from a Dirichlet
Problem
We are currently using a constrained simplex for the weekday effect with a normal prior for the mean of each element. It isn't clear how stan implements this under the hood, and would be better to sample directly from a Dirichlet
Thanks for the flag @damonbayer @dylanhmorris