Classification Performance Analysis of Decision Tree-Based Algorithms with Noisy Class Variable

Table 6

Friedman’s average rankings of the algorithms on the data sets with different noise levels, for a level of significance of 0.05.

Algorithm	0% noise	10% noise	20% noise	30% noise
C4.5	2.25	2.45	2.15	2.1
CART	2.65	2.2	2.05	1.95
RF	1.1	1.3	1.75	1.95