Set noise prior variance to current value for linear growth model

R/inst/stan/prophet.stan

@@ -104,7 +104,7 @@ model {
   k ~ normal(0, 5);
   m ~ normal(0, 5);
   delta ~ double_exponential(0, tau);
-  sigma_obs ~ normal(0, 0.1);
+  sigma_obs ~ normal(0, 0.5);
   beta ~ normal(0, sigmas);
 
   // Likelihood

python/stan/unix/prophet.stan

@@ -104,7 +104,7 @@ model {
   k ~ normal(0, 5);
   m ~ normal(0, 5);
   delta ~ double_exponential(0, tau);
-  sigma_obs ~ normal(0, 0.1);
+  sigma_obs ~ normal(0, 0.5);
   beta ~ normal(0, sigmas);
 
   // Likelihood

python/stan/win/prophet.stan

@@ -149,7 +149,7 @@ model {
   k ~ normal(0, 5);
   m ~ normal(0, 5);
   delta ~ double_exponential(0, tau);
-  sigma_obs ~ normal(0, 0.1);
+  sigma_obs ~ normal(0, 0.5);
   beta ~ normal(0, sigmas);
 
   // Likelihood