Generates posterior linear predictions for objects fitted by bnec

Source: R/posterior_epred.R

Generates posterior linear predictions for objects fitted by bnec. object should be of class bayesnecfit or bayesmanecfit.

# S3 method for bayesnecfit
posterior_epred(object, ...)

# S3 method for bayesmanecfit
posterior_epred(object, ...)

Arguments

object

An object of class bayesnecfit or bayesmanecfit.

...

Additional arguments to posterior_epred.

Value

See ?brms:posterior_epred.

Examples

if (FALSE) {
library(bayesnec)
# Uses default `resolution` and `x_range` to generate `newdata` internally
posterior_epred(manec_example)
# Provide user-specified `newdata`
nd_ <- data.frame(x = seq(0, 3, length.out = 200))
ppreds <- posterior_epred(manec_example, ecx_val = 50, newdata = nd_,
                            make_newdata = FALSE)
ncol(ppreds) == 200 # cols are x, rows are iterations
# Predictions for raw input data
nec4param <- pull_out(manec_example, model = "nec4param")
preds <- posterior_epred(nec4param, make_newdata = FALSE)
x <- pull_brmsfit(nec4param)$data$x
plot(sort(x), preds[1, order(x)], type = "l", col = alpha("black", 0.1),
     ylim = c(-6, 3))
for (i in seq_len(nrow(preds))[-1]) {
  lines(sort(x), preds[i, order(x)], type = "l", col = alpha("black", 0.1))
}
}