Table 7

Average and standard deviation of 10-fold CV accuracy from 10 runs
BMSF- SVM GeneSrF- SVM BMSF- NB GeneSrF- NB BMSF- LDA GeneSrF- LDA BMSF- QDA GeneSrF- QDA
CNS 94.11(1.38) 88.23(0) 92.05(2.79) 91.17(0) 96.47(1.86) 83.23(3.93) 95.88(2.05) NA(NA)
Colon 94.35(2.04) 78.22(2.18) 87.41(1.98) 86.93(1.93) 87.41(1.48) 81.45(0.85) 89.51(1.9) 81.12(1.08)
DLBCL 98.57(1.29) 88.44(2.48) 88.96(0.68) 88.57(0.54) 96.23(0.41) 85.71(1.36) 94.15(1.1) 90(0.62)
GCM 98.07(0.61) 92.07(0.9) 87.42(0.76) 84.17(0.24) 91.14(0.66) 77.53(1.41) 90.25(0.93) NA(NA)
Leukemia 98.33(0.58) 96.66(0.97) 96.25(0.93) 94.58(0.43) 98.33(0.58) 93.19(0.43) 97.63(0.67) 94.58(0.78)
Lung 99.11(0.64) 98.95(0.4) 98.39(0.31) 98.34(0) 97.79(0.26) 98.34(0) 97.56(0.46) 98.34(0.26)
Pros1 96.76(0.8) 92.64(1.24) 89.6(0.82) 93.33(0.62) 95.49(0.5) 90.98(1.01) 93.23(0.85) 91.56(0.68)
Pros2 97.38(1.2) 83.75(2.27) 90(1.17) 85.11(1.46) 95.34(1.13) 85.45(0.71) 90.11(2.27) 83.63(1.09)
Pros3 98.48(1.59) 93.93(2.47) 99.69(0.95) 99.69(0.95) 96.66(0.95) 93.93(0) 100(0) 96.66(1.72)

BMSF and random forest (GeneSrF) are used for gene selection and SVM, NB, LDA, QDA are used to build model with the training data and predict the class of test data.

Zhang et al.

Zhang et al. BMC Bioinformatics 2012 13:298   doi:10.1186/1471-2105-13-298

Open Data