Skip to contents

Use sanity_multi() as the fast first screen after fitting. It prints, and returns invisibly, pass / warn status flags covering optimiser convergence, max gradient component, Hessian definiteness, parameter standard errors, and identifiability of the latent() loadings. For report tables use check_gllvmTMB(); for a broader human-readable summary use gllvmTMB_diagnose().

Usage

sanity_multi(object, gradient_thresh = 0.01, se_thresh = 100)

Arguments

object

A fit returned by gllvmTMB().

gradient_thresh

Maximum allowed absolute gradient component. Default 0.01.

se_thresh

Threshold above which a parameter SE is flagged as suspiciously large. Default 100.

Value

Invisibly a list of diagnostic results.

Details

Scope: fast numerical and loading-shape checks for fitted models. A PASS here does not prove interval calibration or latent-rank identifiability; use target-explicit known-DGP simulation studies for those heavier questions.