A New Robust Classifier on Noise Domains: Bagging of Credal C4.5 Trees

Table 12

values of the Nemenyi test about the accuracy on data sets with 20% of added noise.

	Algorithms
21	BA-C4.5-U versus BA-CC4.5	0
20	BA-C4.5-U versus BA-CDT	0
19	BA-CC4.5 versus RF	0
18	BA-C4.5-U versus BA-CC4.5-U	0
17	BA-C4.5-U versus BA-C4.5	0
16	BA-CDT-U versus BA-CC4.5	0.000001
15	BA-CDT versus RF	0.000001
14	BA-CC4.5-U versus RF	0.000001
13	BA-CDT-U versus BA-CDT	0.000038
12	BA-C4.5 versus RF	0.000038
11	BA-CDT-U versus BA-CC4.5-U	0.000046
10	BA-CDT-U versus BA-C4.5	0.001194
9	BA-C4.5-U versus BA-CDT-U	0.07102
8	BA-C4.5 versus BA-CC4.5	0.074716
7	BA-CC4.5-U versus BA-CC4.5	0.342638
6	BA-C4.5-U versus RF	0.354539
5	BA-CDT versus BA-CC4.5	0.366699
4	BA-C4.5 versus BA-CDT	0.379114
3	BA-CDT-U versus RF	0.379114
2	BA-CC4.5-U versus BA-C4.5	0.40471
1	BA-CC4.5-U versus BA-CDT	0.963078