Table 4 The average accuracy on 5 sets of parameters and six ranking methods on the NCI60 data. No perfect predictors are found by the algorithm. The highest accuracy is written in bold.
Population size | Feature size | The accuracy of different rank methods on all dataset (LOOCV) [%] | |||||
|---|---|---|---|---|---|---|---|
R1 | R2 | R3 | R4 | R5 | R6 | ||
10 | 30 | 66.72 | 63.77 | 60.00 | 54.43 | 62.78 | 69.34 |
50 | 67.86 | 62.78 | 61.63 | 52.62 | 62.62 | 65.90 | |
30 | 30 | 76.23 | 72.29 | 72.02 | 65.90 | 74.26 | 75.41 |
50 | 73.44 | 72.46 | 71.15 | 63.11 | 73.44 | 73.93 | |
50 | 50 | 75.08 | 72.29 | 71.96 | 71.97 | 73.77 | 74.16 |