Tune.spls measure issues

Support
1

Dear mixOmics’ team,
first of all thanks for this really nice package, if been working with it for months now and I am really impressed by your work.

Something I just noticed is that in the vignette where the parameter tuning of the sPLS function is descriped (Chapter 5 Projection to Latent Structure (PLS) | mixOmics vignette) the used measure is ‘MAE’, but when I use the exact same code as in the example I get the following error message:

data(nutrimouse)
X2 <- nutrimouse$gene
Y2 <- nutrimouse$lipid
MyResult.pls <- pls(X2,Y2, ncomp = 4)
set.seed(30) # for reproducbility in this vignette, otherwise increase nrepeat
perf.pls <- perf(MyResult.pls, validation = "Mfold", folds = 5,
                 progressBar = FALSE, nrepeat = 10)
list.keepX <- c(2:10, 15, 20)
# tuning based on MAE
set.seed(30) # for reproducbility in this vignette, otherwise increase nrepeat
tune.spls.MAE <- tune.spls(X2, Y2, ncomp = 3,
                           test.keepX = list.keepX,
                           validation = "Mfold", folds = 5,
                           nrepeat = 10, progressBar = FALSE,
                           measure = 'MAE')

Error in match.arg(measure, choices = c("cor", "RSS")) : 
  'arg' should be one of “cor”, “RSS”

I also checked the documentation of the tune.spls function (tune.spls function - RDocumentation) and there it also says that the possible measures are ‘MAE’, ‘MSE’, … and not ‘cor’ or ‘RSS’. What am I missing here?

I am using mixOmics version 6.22.0.

Best regards,
Katharina

2

You’re missing nothing - its just that the vignette is really quite out of date. I’d recommend the website as the place to go - it has the most up to date resources.

Apologies about the inconsistency. This issue has been on my radar for a while but is low on the priority list sadly

3

Ok, good to know. Thanks!