Closed AlainZuur closed 3 months ago
Great stuff Alain! Also related to one of my recent commits.
Woops, closed this as I mistook it for the bug report in #166.
For the emmeans function I wrote in the discussion I needed both the original matrix and the design matrix for the predictors in the LVs. So, model$lv.X is not the original data.frame passed to the function, and model$lv.X.design now holds the design matrix instead. It's cleaner to separate things that way, and corresponds to the organisation of matrices for "main" X.
I'll make a note in "news" of this change.
Am I correct that in the development version, M6$lv.X is not a matrix anymore? I think in the CRAN version it was. No big deal...I just add:
X <- as.matrix(M6$lv.X)
Alain