PSIS-LOO influence gallery¶

This is a supplementary Python visual guide. It does not change the frozen 59/59 canonical article parity. Every image is generated deterministically from package functions.

Pointwise ELPD-LOO generated by gp3bayespy — **Pointwise ELPD-LOO.**
`plot_loo_pointwise_elpd()`

Pareto-k versus ELPD generated by gp3bayespy — **Pareto-k versus ELPD.**
`plot_loo_pareto_vs_elpd()`

Ranked PSIS-LOO influence generated by gp3bayespy — **Ranked PSIS-LOO influence.**
`plot_loo_influence_rank()`

Grouped PSIS-LOO influence generated by gp3bayespy — **Grouped PSIS-LOO influence.**
`plot_loo_group_influence()`

Grouped predictive contribution generated by gp3bayespy — **Grouped predictive contribution.**
`plot_loo_group_elpd()`

Interpretation boundary

Visual diagnostics remain descriptive evidence. They do not automatically certify adequacy, exclude observations, or select a model.

Back to the complete plot gallery.