Define how to draw from posterior distribution

vmc uses the R package tidybayes to draw from posterior distributions. For posterior predictive distributions, vmc uses tidybayes::predicted_draws(); for posterior distribution of push-forward transformation, vmc uses tidybayes::linpred_draws().

mc_draw(
  distribution = "prediction",
  newdata = NULL,
  extract_function = NULL,
  response_var = NULL,
  ndraws = 500,
  transform = TRUE,
  ...
)

Arguments

distribution: Which distribution to draw from. The options include "prediction" and and push-forward transformations (e.g. mu, sigma, and phi). For example, if the model is normal family and distribution = "prediction", we draws from posterior predictive distribution; if distribution = "mu", we draws from the distribution of linear/link-level predictor (i.e. push-forward transformations).
newdata: Data frame to generate predictions from, or NULL to reuse the data frame used to fit model, i.e. replicated predictive distribution.
extract_function: The function used to draw from model's posterior distributions. Default to be NULL. If extract_function is NULL, then vmc will use tidybayes::predicted_draws() for distribution = "prediction" and use tidybayes::linpred_draws() for other distributions of linear/link-level predictors. If extract_function is not NULL, it should be a function that takes model, newdata, and ndraws as inputs and output a data frame that sampled from model on newdata with at least a prediction column (the draws from the model's distribution), a .row column (a factor grouping rows from the input newdata), and a .draw column (a unique index corresponding to each draw from the distribution).
response_var: A string for the response variable in model. Default to be NULL. If NULL, vmc will infer the response variable from model$formula.
ndraws: The number of draws to return, or NULL to return all draws.
transform: Whether to transform the distribution by its link function. Default to be TRUE.
...: Augments passed to extract_function. If extract_function is NULL, then ... will be passed to tidybayes::predicted_draws() or tidybayes::linpred_draws()

Examples

library(tidybayes)
library(dplyr)
library(modelr)

mcplot(mpg_model) +
  mc_draw()
#> Warning: No shared levels found between `names(values)` of the manual scale and the
#> data's fill values.


mcplot(mpg_model) +
  mc_draw("mu")
#> Warning: No shared levels found between `names(values)` of the manual scale and the
#> data's fill values.


mcplot(mpg_model) +
  mc_draw("mu", data_grid(mpg_model$data, disp, vs, am))
#> Warning: No shared levels found between `names(values)` of the manual scale and the
#> data's fill values.


epred_draws_mu = function(model, newdata, ...) {
  epred_draws(model, newdata, dpar = "mu", ...) %>%
    mutate(prediction = mu)
}

mcplot(mpg_model) +
  mc_draw(extract_function = epred_draws_mu)
#> Warning: No shared levels found between `names(values)` of the manual scale and the
#> data's fill values.