Computes per-parameter improved Rhat and bulk / tail effective sample size
by delegating to the engine's tulpa::diagnostics(), the single source of
truth for convergence diagnostics in the tulpa ecosystem. Works with all
backends.
Usage
# S3 method for class 'ratiod_fit'
diagnostics(fit, pars = NULL, ...)
# S3 method for class 'ratiod_diagnostics'
print(x, ...)Details
Rhat is the improved value of Vehtari et al. (2021): the maximum of rank-normalized split-Rhat and folded split-Rhat. Split-Rhat is defined for a single chain (the chain is split in half), so single-chain fits also get a value.
Examples
# \donttest{
sim <- sim_ratiod(n = 100, family = ratiod_poisson_gamma(), seed = 1)
fit <- tratio(y_num | y_denom ~ x1 + (1 | group), data = sim$data,
family = ratiod_poisson_gamma(),
control = list(iter = 1000, chains = 2, seed = 1))
diagnostics(fit)
#> MCMC Diagnostics
#> ================
#>
#> parameter rhat ess_bulk ess_tail
#> beta_num[1] 1.042 128 149
#> beta_num[2] 1.002 406 334
#> beta_denom[1] 1.044 130 147
#> beta_denom[2] 1.013 186 147
#> sigma_re 1.013 157 211
#> re[1] 1.030 150 181
#> re[2] 1.045 143 149
#> re[3] 1.015 165 159
#> re[4] 1.031 135 205
#> re[5] 1.034 132 164
#> re[6] 1.033 131 178
#> re[7] 1.039 136 134
#> re[8] 1.027 142 160
#> re[9] 1.029 133 183
#> re[10] 1.033 137 126
#> shape 1.010 344 336
#>
#> Warning: 14 parameter(s) have Rhat > 1.01
#> Warning: 16 parameter(s) have ESS < 400
# }