Skip to main content
. 2019 May 22;35(24):5257–5263. doi: 10.1093/bioinformatics/btz421

Table 1.

PCA-LDA fitting accuracy for training and cross-validation (CV) varying with the sample selection method (RS: random selection; KS: Kennard-Stone; MLM: Morais-Lima-Martin) applied in datasets 1–6

Dataset Sample selection method Training accuracy (%) CV accuracy (%)
1 RS 84 83
KS 80 80
MLM 83 82
2 RS 85 83
KS 81 80
MLM 82 77
3 RS 86 84
KS 83 82
MLM 84 80
4 RS 92 91
KS 90 90
MLM 93 90
5 RS 93 93
KS 89 88
MLM 91 88
6 RS 74 72
KS 75 72
MLM 76 75