Table 4.
Performance comparison on the simulated data sets
| Method | SNR | 1/0.6 | 1/1 | 1/2 |
|---|---|---|---|---|
| EP-cor | 0.579 (0.009) | 0.497 (0.012) | 0.368 (0.010) | |
| Step-LR | NT | 3.8 (0.42) | 3.8 (0.42) | 3.3 (0.67) |
| NF | 4.9 (2.64) | 8.6 (3.63) | 7.5 (3.54) | |
| EP-cor | 0.579 (0.011) | 0.498 (0.013) | 0.376 (0.018) | |
| MARS | NT | 3.9 (0.32) | 3.5 (0.53) | 3.3 (0.48) |
| NF | 4.4 (1.78) | 4.2 (1.81) | 4.6 (2.80) | |
| EP-cor | 0.603 (0.007) | 0.528 (0.010) | 0.405 (0.011) | |
| Boost | NT | 3.9 (0.32) | 3.7 (0.48) | 3.5 (0.53) |
| NF | 0.6 (0.70) | 2.4 (1.58) | 5.9 (1.85) | |
| EP-cor | 0.636 (0.009) | 0.551 (0.007) | 0.416 (0.006) | |
| BART | NT | 4.0 (0.00) | 3.6 (0.70) | 3.5 (0.71) |
| NF | 1.8 (1.40) | 2.5 (1.18) | 2.5 (1.43) |
Reported are the averages (standard deviations in the parentheses) over 10 independent data sets. NT and NF are the numbers of true and false motifs identified, respectively.