Compute a Bayesian version of R-squared for MixMustR mixture models

mixmustr_bayes_R2(modfit, summary = TRUE, ...)

Arguments

modfit: A fitted Stan model object containing posterior draws. Typically produced by function run_mixmustr_models.
summary: Should summary statistics be returned instead of the raw values? Default is TRUE.
...: Additional unused arguments.

Value

If summary = TRUE, a (J + 1) x 4 matrix is returned (J + 1 = number of sampled + 1 unsampled source). Columns contain source names and summary statistics (mean +/- 95% HDI) of the Bayesian R-squared values. If summary = FALSE, the posterior draws of the Bayesian R-squared values are returned in an S x (J + 1) matrix (S is the number of draws).

Details

For an introduction to the approach, see Gelman et al. (2019) and https://github.com/jgabry/bayes_R2/.

References

Andrew Gelman, Ben Goodrich, Jonah Gabry & Aki Vehtari. (2019). R-squared for Bayesian regression models, The American Statistician, 73(3):307-309. 10.1080/00031305.2018.1549100

Examples

if (FALSE) { # \dontrun{
library(MixMustR)
# mixmustr_models[6, ] runs quickest
model_fits <- run_mixmustr_models(
  mixmustr_models[6, ], synthetic_df_convergent, tracer_parameters,
  sigma_ln_rho = 0.1, iter = 1e4, warmup = 5e3, chains = 4, cores = 4
)
# Now compute R2, using data stream 2 as point of comparison
mixmustr_bayes_R2(
  model_fits[[1]]$model, data_streams_list = synthetic_df_convergent,
  target = "df_stream_2", order_ref = tracer_parameters$mus$source
)
} # }