Run pipeline deriving scoring rule for PDD — run_scoring_rule

Using different assumed PDD prevalences and different models, computes a set of ROC analyses and derives decision rules for the best threshold identified by each.

Usage

run_scoring_rule_pipeline(
  y_obs,
  model,
  scaling,
  linear = TRUE,
  ...,
  nms = NULL
)

Arguments

y_obs: Observed outcome values.
model: Model object used for prediction (needs to be Bayesian and ammenable to functions from the {posterior}package.
scaling: A scaling table with columns "x" for predictor name, "M" for mean and "SD" for standard deviation..
linear: Should predictions from the linear model (TRUE) or on the response scale (FALSE) be used?
...: Other parameters going either to run_roc if prevs, to extract_coefficients if stat, to or to compute_predictions if seed.
nms: Optional coefficient names, including intercepts. Double check they are in correct order!

Value

description