add optional lower bound on diag

R/fit.R

@@ -16,6 +16,8 @@
 #' @param varss Boolean, whether to fit a VARSS model (defaults to true) or VAR model
 #' @param b_diag A 2 element list specifying the mean and sd of a normal prior on the diagonal elements.
 #' These elements are `mu` (default= 0.7) and `sd` (default = 1)
+#' @param b_diag_min The minimum value of the B diagonal, defaults to 0. Setting this lower,
+#' e.g. -1, allows for oscillations
 #' @param b_offdiag A 2 element list specifying the mean and sd of a normal prior on the off-diagonal elements.
 #' These elements are `mu` (default= 0) and `sd` (default = 1)
 #' @param off_diag_prior Character string denoting which prior to use for
@@ -130,6 +132,7 @@ fit <- function(data,
                 est_trend = FALSE,
                 varss = TRUE,
                 b_diag = list(mu = 0.7, sd =1),
+                b_diag_min = 0,
                 b_offdiag = list(mu = 0, sd = 1),
                 off_diag_prior = c("normal", "student-t", "laplace", "hs","normal_pp"),
                 sigma_pro = list(mu = 0, sd = 1),
@@ -276,6 +279,7 @@ fit <- function(data,
     b_sd = b_offdiag$sd,
     b_mu_diag = b_diag$mu,
     b_sd_diag = b_diag$sd,
+    b_diag_min = b_diag_min,
     nu_known = nu_known,
     sigma_proc_sd = sigma_pro$sd,
     sigma_proc_mu = sigma_pro$mu,

inst/stan/varlasso.stan

@@ -32,6 +32,7 @@ data {
   real yy[n_pos];       // data
   //real b_mu[n_off]; // prior on mean of offdiagonal
   real b_sd[n_off]; // prior on sd offdiagonal
+  real b_diag_min; // minimum value of estimated B diag
   real b_mu_diag[n_spp];// prior on mean of diagonal
   real b_sd_diag[n_spp];// prior on sd of diagonal
   real fixed_r[n_spp]; // fixed_r is optional, will be all 0s if estimated
@@ -106,7 +107,7 @@ parameters {
   real<lower=0> sigma_obs[est_sigma_obs * n_r];
   vector<lower=0>[est_sigma_scale] sigma_scale; // variance for shrinkage / hierarchical B off diags
   vector[n_off] B_z;  // off-diags of B, in normal (0,1) space
-  vector<lower=0,upper=1>[n_spp] Bdiag;   // diag of B
+  vector<lower=b_diag_min,upper=1>[n_spp] Bdiag;   // diag of B
   vector[n_spp] x0;
   vector[est_trend*n_spp] U;
   vector<lower=2>[est_nu] nu; // student-t df parameters for Student-t priors