Posterior Prediction for idealstan objects

Description

This function will draw from the posterior distribution, whether in terms of the outcome (prediction) or to produce the log-likelihood values.

This function can also produce either distribution of the outcomes (i.e., predictions) or the log-likelihood values of the posterior (set option type to ‘log_lik’. For more information, see the package vignette How to Evaluate Models.

You can then use functions such as id_plot_ppc to see how well the model does returning the correct number of categories in the score/vote matrix. Also see help(“posterior_predict”, package = “rstanarm”)

Usage

## S4 method for signature 'idealstan'
id_post_pred(
  object,
  draws = 100,
  output = "observed",
  type = "predict",
  covar = "person",
  sample_scores = NULL,
  item_subset = NULL,
  use_cores = 1,
  use_chain = NULL,
  newdata = NULL,
  ...
)

Arguments

object	A fitted idealstan object
draws	The number of draws to use from the total number of posterior draws (default is 100). Set to "all" to use all draws in the chains. For reproducibility, you can also pass a vector of specific draws to use.
output	If the model has an unbounded outcome (Poisson, continuous, etc.), then specify whether to show the ‘observed’ data (the default) or the binary output ‘missing’ showing whether an observation was predicted as missing or not
type	Whether to produce posterior predictive values (‘predict’, the default), the posterior expected (average) values (‘epred’), or log-likelihood values (‘log_lik’). See the How to Evaluate Models vignette for more info.
covar	What kind of covariates to include as part of the prediction – either "person" (the default) or "items" if you included predictors for item discriminations.
sample_scores	In addition to reducing the number of posterior draws used to calculate the posterior predictive distribution, which will reduce computational overhead. Only available for calculating predictive distributions, not log-likelihood values.
item_subset	Whether to calculate marginal effects for only a subset of items. Should be item IDs that match the item_id column passed to the id_make function.
use_cores	Number of cores to use for multicore parallel processing with the base R parallel package
newdata	Optional: pass a data frame that must have all of the predictors that were given to the id_make function. Used to generate predictions from person or item covariates on to items.
…	Any other arguments passed on to posterior_predict (currently none available)