2

In partial least squares (PLS), I have multiple variables in X and only one variable in Y. If I only choose one PLS component to use for the PLS model, can I assume that this PLS component is in the same direction as the single Y variable?

As shown in the picture below, if there are 2 variables in Y, and 2 PLS components are selected, the below picture is the "loading scatter plot" (or "biplot"?). So either variable "C" or "M" is not in the same direction as PLS component 1 or 2.

So I wonder if there is only 1 variable in Y, and only 1 PLS component is used for the model, then if this Y variable will fall onto the PLS component direction?

enter image description here

amoeba
  • 104,745
lanselibai
  • 507
  • 2
  • 6
  • 14
  • I think the correlation will be indicated by the "percentage". For example, if a model can explain 78%, then I think this 78% is the same direction as the Y variable. – lanselibai Apr 23 '16 at 10:33
  • 1
    Thanks for the claryfing update. It's much clearer now. +1 (and I voted to reopen) – amoeba Apr 24 '16 at 16:55

0 Answers0