Plot predictive means against generalised residuals — plot_predicted

Uses influ_residuals(), never native Pearson or deviance residuals. The horizontal coordinate is the predictive mean from the same simulations, preserving their response component and random-effect conditioning. Pass a precomputed result to avoid repeating simulation when styling plots.

Usage

plot_predicted_residuals(fit, trend = "loess", type = "quantile", ...)

Arguments

fit: A supported fitted model or an influ_residuals object.
trend: One of "loess", "lm", "linear", or "none".
type: "quantile", "generalised", or "generalized" (equivalent).
...: Calculation options passed to influ_residuals(). Not accepted for a precomputed object. Complete brms fits are needed for calculation; existing posterior draws are used without fitting or running MCMC.

Value

A ggplot with a residual_metadata attribute describing the target.