gmvjoint version 0.3.0

• Extensive rewrites to workhorse function joint; some computational efficiency gained 'within' the EM algorithm.
• This change also involved now allowing users to supply dispersion models for certain families.
• Removed internal functions which created data matrices. These were slow, and moreover completely redundant since glmmTMB creates all required matrices in the initial condition stage.
• Added bootstrapping function. The approximated SEs returned by default are known to be underestimated, so this was added as an alternative route, or merely a means of comparison.
• Added xtable S3 method.
• Removed bootAUC, since it was kind of problematic, and I wasn't sure how best to approach fixing it at this current time.
• Sprucing up of S3 methods; plotting; print-outs etc.

gmvjoint v0.2.0

Second CRAN release!

Lots of new stuff:

• Dynamic predictions and ROC/AUC functions.
• Other model comparison stuff (fairly skeletal).
• More S3 methods of objects of class joint.
• More flexible calls to joint allowed. And more convergence criterion added.
• Fixes to quadrature for binomial family.
• Facility to generate conditional distributions of random effects by sampling from a joint model fit.

gmvjoint v0.1.0.1

CRAN first release.

gmvjoint v0.1.0

First public release