Skip to main content

Table 2 95%-confidence intervals for the differences in prediction errors.

From: Instance-based concept learning from multiclass DNA microarray data

   1-NN 3-NN 5-NN SVM DT MLP
NCI60 k-NN 0 1.30 ± 0.09 - 1.40 ± 0.16 2.0 ± 0.21 1.60 ± 0.16
  1-NN - 1.30 ± 0.09 - 1.40 ± 0.16 2.0 ± 0.21 1.60 ± 0.16
  3-NN - - - 2.30 ± 0.16 1.30 ± 0.22 1.10 ± 0.15
  SVM - - - - 2.60 ± 0.23 2.60 ± 0.21
  DT - - - - - 1.80 ± 0.21
ALL k-NN 2.60 ± 0.05 0.90 ± 0.02 0.90 ± 0.02 2.90 ± 0.06 10.10 ± 0.06 8.80 ± 0.08
  1-NN - 2.90 ± 0.05 2.90 ± 0.05 1.50 ± 0.03 9.10 ± 0.03 7.80 ± 0.10
  3-NN - - 0 3.20 ± 0.06 10.0 ± 0.06 8.70 ± 0.08
  5-NN - - - 3.20 ± 0.06 10.0 ± 0.06 8.70 ± 0.08
  SVM - - - - 9.80 ± 0.02 8.30 ± 0.09
  DT - - - - - 5.30 ± 0.10
GCM k-NN 0.40 ± 0.02 2.20 ± 0.07 2.20 ± 0.07 2.10 ± 0.08 4.60 ± 0.05 9.30 ± 0.09
  1-NN - 2.20 ± 0.06 2.20 ± 0.06 2.10 ± 0.08 4.80 ± 0.06 9.50 ± 0.09
  3-NN - - 0.00 2.90 ± 0.09 3.80 ± 0.09 7.90 ± 0.09
  5-NN - - - 2.90 ± 0.09 3.80 ± 0.09 7.90 ± 0.09
  SVM - - - - 5.70 ± 0.12 10.00 ± 0.13
  DT - - - - - 4.70 ± 0.11