Skip to main content

Table 2 Estimated error rates (ATE) and standard deviations for an increasing number of features on colon cancer data (average over 50 experiments), for each of the ranking methods.

From: Entropy-based gene ranking without selection bias for the predictive classification of microarray data

# genes E-RFE RFE SQRT-RFE CC TT
10 20.9 ± 10.0 22.0 ± 11.0 19.5 ± 9.8 20.4 ± 10.6 21.3 ± 9.4
20 19.9 ± 8.9 20.0 ± 9.1 19.3 ± 9.5 20.8 ± 9.6 19.2 ± 10.5
50 18.5 ± 8.5 18.4 ± 9.0 17.2 ± 8.1 20.7 ± 8.1 21.2 ± 9.3
100 16.8 ± 7.8 17.3 ± 8.1 17.2 ± 8.2 21.5 ± 9.6 22.0 ± 10.5
300 17.2 ± 8.2 17.6 ± 8.1 17.6 ± 8.2 19.7 ± 10.2 19.9 ± 9.5
500 17.9 ± 7.9 17.9 ± 7.9 17.5 ± 7.4 17.7 ± 8.3 17.5 ± 8.6
1000 16.9 ± 7.5 16.9 ± 7.4 16.8 ± 7.3 17.1 ± 7.5 17.5 ± 7.6
2000 All features: 16.0 ± 7.5