library(dplyr) library(rstan) sim_panel_ri_far1 <- read.csv('random_ar1.csv') sigma_u <- 1 sigma_e <- 3 S <- 50 T <- 100 P = 1 beta = c(1) delta = rnorm(S, .4, .15) y <- matrix(0, T, S) y_lag <- matrix(0, T, S) for (s in 1:S) { y[,s] <- filter(sim_panel_ri_far1, individual == s)$y y_lag[2:T,s] <- filter(sim_panel_ri_far1, individual == s)$lagged_y[2:T] } X_list <- lapply(1:S, function(x) X) #mod <- stan_model('dynamic-panel_w_predictors.stan') stan_data <- list(S = S, T = T, P = P, y = y, y_lag = y_lag, X = X_list) fit.ri.rar1 <- sampling(dynamic_panel_wbeta_ris_ar1, data = stan_data, iter = 2000, cores = 4, chains = 4, control = list(adapt_delta = 0.9)) library(bayesplot) mcmc_recover_intervals(as.matrix(fit.ri.rar1, pars = 'delta'), true = delta)