Help with InvertY.cpp message?

[jacobenglert]

Hi there,

I am using the hybrid sampler from this package for a negative binomial regression model. Occasionally, I will get the following message: "InvertY.cpp, v_eval: reached max_iter". Given it is not an error or warning, does this warrant concern? It appears to only occur when the SP approximation is used, so my only thought is to default to another method when I encounter the message.

Thanks!

[jwindle]

I have to look back at this, but I think you may be able use rpg.devroye or rpg.gamma directly. For rpg.devroye you need an integer shape parameter. rpg.gamma is an approximation, ultimately. I'll look into exposing the max_iter parameter so it is possible to increase it.