TODO.md

In bbolker/glmmadmb: Generalized Linear Mixed Models using 'AD Model Builder'

Wish/to-do list for glmmADMB

• formula interface for zero-inflation
• more complex variance models (ASREML syntax)
• Can we add some sort of progress bar or "working ..." message in the absence of voluminous messages (difficult); better isolation of messages to set command-line flags
• Add more examples, tests, vignettes ... especially a smaller (quicker) data set for the main example
• work on predict() method; predictions incl REs
• simulate() method
• save (more) fitted models in inst/ so they can be used easily in examples without doing the fits each time
• speed up (Skaug)
• more robust first pass, weighted SS or quasi-likelihood (Fournier)
• access to profiling
• MCMC fixes: transform correlation parameter
• sort out terms, formula methods to make step etc. work properly (check lme4 implementation, add fixed.only=FALSE option?)
• would making predictions into sdreport variables give (Wald) CIs that included uncertainty in Var-Corr?
• vignette: more on troubleshooting, R2admb intro, importance sample, AGQ, ... ?

bbolker/glmmadmb documentation built on May 11, 2019, 9:29 p.m.