Extract all values of Bayesian models.
# S3 method for stanreg
model_values(model, ci = 0.9, ci_method = "default",
standardize = FALSE, standardize_robust = FALSE, effsize = NULL,
performance_in_table = TRUE, performance_metrics = c("R2",
"R2_adjusted"), parameters_estimate = "median",
parameters_test = c("pd", "rope"), parameters_diagnostic = TRUE,
parameters_priors = TRUE, rope_range = "default", rope_full = TRUE,
...)Object of class lm.
Credible Interval (CI) level. Default to 0.90 (90%).
Standardized coefficients. See model_parameters.
Interpret the standardized parameters using a set of rules. Can be "cohen1988" (default for linear models), "chen2010" (default for logistic models), "sawilowsky2009", NULL, or a custom set of rules.
Add performance metrics in table.
Can be "all" or a list of metrics to be computed (some of c("LOO", "R2", "R2_adj")).
The point-estimate(s) to compute. Can be a character or a list with "median", "mean" or "MAP".
What indices of effect existence to compute. Can be a character or a list with "p_direction", "rope" or "p_map".
Include sampling diagnostic metrics (effective sample, Rhat and MCSE). Effective Sample should be as large as possible, altough for most applications, an effective sample size greater than 1,000 is sufficient for stable estimates (B<U+00FC>rkner, 2017). Rhat should not be larger than 1.1.
Include priors specifications information. If set to true (current rstanarm' default), automatically adjusted priors' scale during fitting will be displayed.
ROPE's lower and higher limits Should be a list of two values (e.g., c(-0.1, 0.1)) or "default". If "default", the bounds are set to x +- 0.1*SD(response).
If TRUE, use the proportion of the entire posterior distribution for the equivalence test. Otherwise, use the proportion of HDI as indicated by the ci argument.
Arguments passed to or from other methods.