Skip to main content
Journal of the Brazilian Computer Society

Table 2 Comparison among the methods J48, Clusk and ClusEM

From: Clus-DTI: improving decision-tree classification with a clustering-based decision-tree induction algorithm

 

Wins Acc

Wins TS

Margin Acc Wins (vs. J48)

Margin TS Wins (vs. J48)

Margin Acc Loss (vs. J48)

Margin TS Loss (vs. J48)

J48

14

2

–

–

–

–

Clusk

13

63

4.40 %

23.4

4.30 %

1.75

ClusEM

10

18

4.20 %

28.2

4.40 %

0.91

  1. Wins Acc—number of times each method has outperformed the other two regarding accuracy
  2. Wins TS—number of times each method has outperformed the other two regarding tree size
  3. Margin Acc Wins (vs. J48)—average difference in accuracy when a Clus-DTI method outperforms J48
  4. Margin TS Wins (vs. J48)—average difference in tree size when a Clus-DTI method outperforms J48
  5. Margin Acc Loss (vs. J48)—average difference in accuracy when a Clus-DTI method is outperformed by J48
  6. Margin TS Loss (vs. J48)—average difference in tree size when a Clus-DTI method is outperformed by J48
Back to article page