Table 1 Best combination of pre-processing and machine learning algorithms after 100 repetitions.

From: Potential of spectroscopic analyses for non-destructive estimation of tea quality-related metabolites in fresh new leaves

Phenotypes

Most selected pre-processing and models

Frequency

(100 repeat−1)

Catechins

GC

DT-cubist

32

C

SNV-cubist

17

CG

DT-RF, OR-cubist

8

EC

DT-cubist

18

ECG

DT-cubist

23

EGC

DT-cubist

65

EGCG

DT-cubist

37

EGCG3Me

DT-cubist

23

Total

DT-cubist

42

FAAs

Asp

DT-cubist

23

Glu

DT-cubist

50

Arg

DT-cubist

25

Thea

DT-cubist

27

Total

DT-cubist

34

Caffeine

DT-cubist

20

  1. Combination of pre-processing and machine learning algorithms were evaluated based on RPD values.