Figure 3. Predictive error sum of squares in model training for monkey B.

The blue line and green line show predictive error sum of squares (PRESS) and R² values, respectively, for different numbers of latent variables used in the PLS model. The optimal number of 20 is denoted with the red dotted line.