Skip to main content

Table 6 Comparison of the classification accuracies for HBSA-SVM(Biased), HBSA-SVM(Unbiased) and HBSA-KNN methods with the top-ranked genes

From: Finding minimum gene subsets with heuristic breadth-first search algorithm for robust tumor classification

Dataset HBSA-SVM (Biased) HBSA-SVM (Unbiased) HBSA-KNN
#TG* Acc.% #TG* Acc.% #TG* Acc.%
Leukemia 2 88.46 3 82.69 2 84.62
3 92.31 5 86.54 5 90.38
71 96.15 (H) 15 92.31 (H) 24 96.15 (H)
DLBCL 2 95.24 2 80.95 2 80.95
3 100 (H) 9 95.24 (H) 3 90.48 (H)
Prostate 1 94.12 1 91.18 (H) 2 88.24 (H)
2 97.06 (H) 4 88.24 5 85.29
SRBCT 4 75 4 70 3 75
5 95 5 90 4 95
24 100 (H) 28 100 (H) 5 100 (H)
ALL 6 97 6 96 5 94
7 96 10 97 9 96
112 100 (H) 111 99 (H) 85 100 (H)
Colon 2 75 2 65 3 70 (H)
3 80 (H) 7 70 4 80 (H)
  4 80 15 80(H) 7 80
  1. *‘#TG’ denotes the number of top-ranked genes. Note that the accuracy labeled by ‘H’ denotes the highest accuracy and the number of the corresponding top-ranked genes denotes the minimal number with the highest accuracy.