Method Training data (R2) % Training data (sd ) MSEP MSEP (sd) Test data  (R2)%      RR 48.1 1.979 1.30 0.031  36.1 LASSO 61.8 5.242 1.24 0.033 41.4 EN 65.4 3.804 1.21 0.035 44.1 PCR 61.8 7.180 1.29 0.032 37.2 PLS 60.0 8.927 1.32 0.037 35.9 SPLS 36.5 11.593 1.31 0.044 36. 36.536.5     36.5 RF 25.0 4.653 1.27 0.032 40.5 SVM 79.7 6.713 1.37 0.032 30.8

Table 1: Comparison of the eight multivariate regression methods based on R2 for training data, standard deviation of R2 (sd) for training data, MSEP, sd of MSEP and R2 for test data set. R2 for training data, MSEP and R2 for test data are the mean values of the double cross validation scheme for 10 different divisions with 100 runs and then averaged.