mvgam (version 1.1.4)

pp_check.mvgam: Posterior Predictive Checks for mvgam models

Description

Perform unconditional posterior predictive checks with the help of the bayesplot package.

Usage

# S3 method for mvgam
pp_check(
  object,
  type,
  ndraws = NULL,
  prefix = c("ppc", "ppd"),
  group = NULL,
  x = NULL,
  newdata = NULL,
  ...
)

Value

A ggplot object that can be further customized using the ggplot2 package.

Arguments

object

An object of class mvgam.

type

Type of the ppc plot as given by a character string. See PPC for an overview of currently supported types. You may also use an invalid type (e.g. type = "xyz") to get a list of supported types in the resulting error message.

ndraws

Positive integer indicating how many posterior draws should be used. If NULL all draws are used. If not specified, the number of posterior draws is chosen automatically. Ignored if draw_ids is not NULL.

prefix

The prefix of the bayesplot function to be applied. Either `"ppc"` (posterior predictive check; the default) or `"ppd"` (posterior predictive distribution), the latter being the same as the former except that the observed data is not shown for `"ppd"`.

group

Optional name of a factor variable in the model by which to stratify the ppc plot. This argument is required for ppc *_grouped types and ignored otherwise.

x

Optional name of a variable in the model. Only used for ppc types having an x argument and ignored otherwise.

newdata

Optional dataframe or list of test data containing the variables included in the linear predictor of formula. If not supplied, predictions are generated for the original observations used for the model fit.

...

Further arguments passed to predict.mvgam as well as to the PPC function specified in type.

Author

Nicholas J Clark

Details

Unlike the conditional posterior checks provided by ppc, This function computes unconditional posterior predictive checks (i.e. it generates predictions for fake data without considering the true observations associated with those fake data). For a detailed explanation of each of the ppc functions, see the PPC documentation of the bayesplot package.

See Also

ppc, predict.mvgam

Examples

# \donttest{
simdat <- sim_mvgam(seasonality = 'hierarchical')
mod <- mvgam(y ~ series +
              s(season, bs = 'cc', k = 6) +
              s(season, series, bs = 'fs', k = 4),
            data = simdat$data_train,
            chains = 2,
            silent = 2)

# Use pp_check(mod, type = "xyz") for a list of available plot types

# Default is a density overlay for all observations
pp_check(mod)

# Rootograms particularly useful for count data
pp_check(mod, type = "rootogram")

# Grouping plots by series is useful
pp_check(mod, type = "bars_grouped",
        group = "series", ndraws = 50)
pp_check(mod, type = "ecdf_overlay_grouped",
        group = "series", ndraws = 50)
pp_check(mod, type = "stat_freqpoly_grouped",
        group = "series", ndraws = 50)

# Custom functions accepted
prop_zero <- function(x) mean(x == 0)
pp_check(mod, type = "stat", stat = "prop_zero")
pp_check(mod, type = "stat_grouped",
        stat = "prop_zero",
        group = "series")

# Some functions accept covariates to set the x-axes
pp_check(mod, x = "season",
        type = "ribbon_grouped",
        prob = 0.5,
        prob_outer = 0.8,
        group = "series")

# Many plots can be made without the observed data
pp_check(mod, prefix = "ppd")
# }