Posterior predictions of vaccination coverage

Draw from the posterior predictive distribution.

# S3 method for icfit
predict(
  X,
  country = NULL,
  vaccine = NULL,
  t = 2,
  ratio_adj = TRUE,
  return_ic = TRUE
)

Arguments

X	object of type icfit or iclist
country	character vector of country codes to predict
vaccine	character vector of vaccine abbreviations to predict
t	Integer of the number of time steps ahead to predict. Default is 2.
ratio_adj	Logical. Should multi-dose vaccines modelled by a ratio be back-transformed to coverage? Default is TRUE.
return_ic	Logical. Should an icfit object be returned?

Value

If return_ic is FALSE, then a data.frame (or a list of data frames) with posterior samples for predictions, labelled by 'country', 'vaccine', and 'time' columns. Returning an 'ic' object will modify object X and add/update with a the 'prediction' element and return the same type as X.

The default behaviour is to back-transform multi-dose vaccines which have been converted to ratios into coverages. This is done by predicting the ratio and multiplying the later dose ratio by the earlier dose's coverage.