Skip to contents
pigauto
0.10.0.9000
Get started
Articles
Knitted vignettes
Get started
Common pitfalls / FAQ
Mixed-type traits
Propagating tree uncertainty
GNN architecture and the math behind pigauto
Methodology
Per-trait benches
Continuous (BM / OU / regime / nonlinear)
Binary (signal strength sweep)
Ordinal (level count sweep)
Count (Poisson / NegBin)
Categorical (K-level sweep)
Proportion (signal sweep)
Zero-inflated counts (zero fraction sweep)
Cross-dataset benches and simulations
AVONET missingness sweep (20 / 50 / 80 %)
Missingness mechanisms (MCAR / MAR / MNAR)
Tree uncertainty (Rubin's rules, 10 posterior trees)
Delhey 5,809-species birds
Multi-observation imputation (CTmax-like)
Covariate effectiveness (simulation sweep)
Reference
Changelog
License
YEAR: 2025 COPYRIGHT HOLDER: Shinichi Nakagawa