Skip to contents

Two-panel diagnostic showing the dual criterion driving \(k^{**}\):

  • Top panel: \(R_v / \hat{SE}^{param}_v\) (predictive revision normalised by parameter SE), with horizontal guide at the threshold c.

  • Bottom panel: \(\hat{D}_v\) (robust cross-cohort dispersion of incremental loss ratio), with horizontal guide at the threshold tau.

Vertical guides mark k_star (dashed) and k_conv (solid). A point falling below both threshold lines passes the joint criterion.

Usage

# S3 method for class 'LRConvergence'
plot(x, theme = c("view", "save", "shiny"), ...)

Arguments

x

An object of class LRConvergence.

theme

String passed to .switch_theme().

...

Additional arguments passed to .switch_theme().

Value

A ggplot object.