Loving the background prediction aspect of PLS-DA analysis.
perf() results, the best results should be with 3 components. However, the background mapping as far as I have managed to see is for 2 components only. It would help our sample mapping and cluster separation to map this in 3D with all 3 components.
Is that an option?
At this stage you will have to keep to a comp 1,2 and comp 1,3. We only have 3D plots for plotIndiv without the background.
Fair enough! Is there any work on this on the cards in the next year or so? Before I try and come up with a work around
No funding so far for a software developer so I dont think we will get around this issue for quite some time!