Weights in manyglm about mvabund HOT 2 OPEN

Roadmap: this is not trivial and will involve several steps:

• manyglm.R: infrastructure for passing in a prior.weights argument and checking, ie no negative, right dimensions ect
  glm.cpp:
• modifying the betaEst function to reflect weighted observations
• modifying the EstIRLS function to reflect weighted observations
  Might also have to do this for the NegBin versions of these as well.

resamptest.h

• modify all family specific functions to accept weights: ie
• llfunc(double yi, double mui, double a)
• devfunc(double yi, double mui, double a)
  maybe need to modify :
• pdf(double yi, double mui, double a)
• cdf(double yi, double mui, double a)
• cdfinv(double u, double mui, double a)
• genRandist(double mui, double a)

Update all calls to glm::regression to have the additional priorWeights Argument, ie in anova and summary functions.

All

• rewriting documentation
• add unit tests

from mvabund.

I just saw this and was wondering whether anything had come of this at this point in time? Thank you!

from mvabund.