brms (version 2.1.0)

pp_average.brmsfit: Posterior predictive samples averaged across models

Description

Compute posterior predictive samples averaged across models. Weighting can be done in various ways, for instance using Akaike weights based on information criteria or marginal likelihoods.

Usage

# S3 method for brmsfit
pp_average(x, ..., weights = "loo", method = c("predict",
  "fitted", "residuals"), newdata = NULL, re_formula = NULL,
  allow_new_levels = FALSE, sample_new_levels = "uncertainty",
  new_objects = list(), incl_autocor = TRUE, resp = NULL,
  nsamples = NULL, sort = FALSE, nug = NULL, summary = TRUE,
  robust = FALSE, probs = c(0.025, 0.975), more_args = NULL,
  control = NULL)
pp_average(x, ...)

Arguments

x

A brmsfit object.

...

More brmsfit objects.

weights

Name of the criterion to compute weights from. Should be one of "loo" (default), "waic", "kfold", or "bridge" (log marginal likelihood). Alternatively, a numeric vector with pre-specified weights.

method

Type of predictions to average. Should be one of "predict" (default), "fitted", or "residuals".

newdata

An optional data.frame for which to evaluate predictions. If NULL (default), the original data of the model is used.

re_formula

formula containing group-level effects to be considered in the prediction. If NULL (default), include all group-level effects; if NA, include no group-level effects.

allow_new_levels

A flag indicating if new levels of group-level effects are allowed (defaults to FALSE). Only relevant if newdata is provided.

sample_new_levels

Indicates how to sample new levels for grouping factors specified in re_formula. This argument is only relevant if newdata is provided and allow_new_levels is set to TRUE. If "uncertainty" (default), include group-level uncertainty in the predictions based on the variation of the existing levels. If "gaussian", sample new levels from the (multivariate) normal distribution implied by the group-level standard deviations and correlations. This options may be useful for conducting Bayesian power analysis. If "old_levels", directly sample new levels from the existing levels.

new_objects

A named list of objects containing new data, which cannot be passed via argument newdata. Currently, only required for objects passed to cor_sar and cor_fixed.

incl_autocor

A flag indicating if ARMA autocorrelation parameters should be included in the predictions. Defaults to TRUE. Setting it to FALSE will not affect other correlation structures such as cor_bsts, or cor_fixed.

resp

Optional names of response variables. If specified, fitted values of these response variables are returned.

nsamples

Positive integer indicating how many posterior samples should be used. If NULL (the default) all samples are used. Ignored if subset is not NULL.

sort

Logical. Only relevant for time series models. Indicating whether to return predicted values in the original order (FALSE; default) or in the order of the time series (TRUE).

nug

Small positive number for Gaussian process terms only. For numerical reasons, the covariance matrix of a Gaussian process might not be positive definite. Adding a very small number to the matrix's diagonal often solves this problem. If NULL (the default), nug is chosen internally.

summary

Should summary statistics (i.e. means, sds, and 95% intervals) be returned instead of the raw values? Default is TRUE.

robust

If FALSE (the default) the mean is used as the measure of central tendency and the standard deviation as the measure of variability. If TRUE, the median and the median absolute deviation (MAD) are applied instead. Only used if summary is TRUE.

probs

The percentiles to be computed by the quantile function. Only used if summary is TRUE.

more_args

Optional list of further arguments passed to the function specified in method.

control

Optional list of further arguments passed to the function specified in weights.

Value

Same as the output of the method specified in argument method.

Examples

# NOT RUN {
# model with 'treat' as predictor
fit1 <- brm(rating ~ treat + period + carry, data = inhaler)
summary(fit1)

# model without 'treat' as predictor
fit2 <- brm(rating ~ period + carry, data = inhaler)
summary(fit2)

# compute model-averaged predicted values
(df <- unique(inhaler[, c("treat", "period", "carry")]))
pp_average(fit1, fit2, newdata = df)

# compute model-averaged fitted values
pp_average(fit1, fit2, method = "fitted", newdata = df)
# }
# NOT RUN {
# }