JenniNiku/gllvm

error with predict.region = TRUE

AlainZuur opened this issue · 1 comments

AlainZuur commented

Development version. Runs fine without the prediction.region. Ran ok in the CRAN version.

ordiplot(M4,

  •      biplot = TRUE,

  •      s.cex = 0.4,

  •      s.colors = grey(0.5),

  •      cex.env = 0.9,

  •      predict.region = TRUE,

  •      col.ellips = "green")

Error in Q[(1:n) + n * (q - 1) + radidx, which(names(fit$TMBfn$par[fit$Hess$incl]) == :
incorrect number of subscripts on matrix

M4 is a GLLVM with 2 constrained latent variables.

No hurries! It is not a thing I use often.

Alain

BertvanderVeen commented

Addressed and solved. Thanks for reporting; caused by one of my recent changes as well.