Closed quang-huynh closed 1 year ago
Thanks for your interest! I think having the ability to flexibly assign error distributions to different observations is one advantage to using a more flexible interface than is available in glm/gam frameworks.
please replace e_i = c(rep(0, 25), rep(1, 175))
and it should work. I'll add an informative error message checking the length of e_i
. I'm closing for now but please report back with any further issue
I have a multi-category model and wanted to specify an error structure
e_i
that is different from the default, where each category has its own variance. However the TMB model crashes and I'm not able to deduce what's going on, asn_e
andObsModel_ez
seem to be correctly specified frommake_data
.Below is a simple reproducible example: