Open cboettig opened 5 years ago
Few notes for me here so I don't forget ... idea would be if we have replicates (k reps) we would have all parameters be hierarchical (a[k] ... a[k] ~ normal(mu, sigma)).
Waiting on Carl and Henry to have MCMC working then will try in Stan. Current version perhaps improved by fewer parameters.
@lizzieinvancouver
Explore the hierarchical estimation of
a
over replicates. (replicates draw their a's from the distribution rather than fixed)