Skip to main content
. 2018 Feb 12;14(3):31. doi: 10.1007/s11306-018-1321-4

Table 4.

Quality of PLS-DA model for the discrimination of two groups when different normalization approaches are applied on the data. The table is reproduced with permission from

(Saccenti 2017)

Method NMC Q 2 DQ 2 AUROC
1 TSA 64 0.02 0.02 0.60
2 IS 77 − 0.33 0.005 0.37
3 UO 57 − 0.04 0.007 0.61
4 PQN 62 − 0.05 − 0.03 0.64
5 Loess 1 0.89 0.92 1
6 Contrast 23 0.41 0.56 0.99
7 Quantile 7 0.63 0.81 0.99
8 Linear 86 − 0.69 − 0.02 0.28
9 Li Wong 64 − 0.25 − 0.03 0.56
10 Spline 7 0.36 0.7 1
11 VSN 0 0.95 0.97 1
12 None 33 0.21 0.25 0.77

NMC Number of misclassifications, DQ2 discriminant Q2, AUROC area under the receiver operating curve. The measures are discussed in (Saccenti 2017)