XY-variates scores plot


I have a question about PLS and its representation. According to mixOmics page (Chapter 5 Projection to Latent Structure (PLS) | mixOmics vignette) we can plot X scores, Y scores (two separated plots) or an average between both components rep.space = ‘XY-variate’.
My question is : Is there a particular interest to plot an average between X and Y scores? or it is just a way ti display data? Do average between X and Y scores have an implication in some further analysis? i.e. prediction? .
When I plot X scores and X-Y scores there are a huge difference. X-Y scores plot separates even more the groups (see exemple bellow).

Thanks very much.


X scores plot

X-Y scores

hi @enzo,

the XY plot enables to combine two plots into one, in a sort of ‘consensus’ plot, but it would only make sense if your X score plot and your Y score plot are very similar (i.e. the components are well correlated across data sets). One way to visualise this is the use the plotArrow().
However, it seems from your plot that you are using a PLS-DA, rather than a PLS? In that case the option XY-variates should not be activated and you should only visualise the X score plot (as the Y score plot shows the obvious, i.e. the two sample groups from the categorical factor Y). This may explain your output.