Table 2.
Predictive Accuracies Obtained by Using Each of the 10 Peak Sets Derived from This Work As Input to SVM Classifiersa
| peak sets | number of selected peaks in each set |
accuracy from 500 test sets |
peaks selected | ||
|---|---|---|---|---|---|
| sensitivity | specificity | accuracy | |||
| Peak Set 1 | 9 | 99.99% | 99.99% | 99.99% | 1580, 1996, 2040, 2187, 2192, 2851, 2893, 4311, 4502 |
| Peak set 2 | 9 | 99.99% | 99.95% | 99.97% | 1580, 1996, 2040, 2187, 2192, 2286, 2851, 4311, 4502 |
| Peak Set 3 | 12 | 99.64% | 100% | 99.79% | 1580, 1800, 1996, 2040, 2187, 2244, 2511, 2851, 4311, 4502, 4516 |
| Peak Set 4 | 11 | 99.99% | 99.98% | 99.99% | 1580, 1996, 2040, 2151, 2187, 2214, 2411, 2711, 2851, 4311, 4502 |
| Peak Set 5 | 11 | 100% | 99.97% | 99.99% | 1580, 1996, 2040, 2187, 2214, 2411,2490, 2851, 2511, 4311, 4502 |
| Peak Set 6 | 9 | 99.99% | 99.96% | 99.98% | 1580, 1996, 2040, 2187, 2192, 2286, 2851, 4311, 4502 |
| Peak Set 7 | 7 | 99.99% | 99.97% | 99.98% | 1580, 1996, 2040, 2187, 2851, 4311, 4502 |
| Peak Set 8 | 9 | 99.90% | 99.90% | 99.90% | 1580, 1800, 1996, 2040, 2187, 2411, 2851, 4311, 4502 |
| Peak Set 9 | 7 | 100% | 99.98% | 99.99% | 1580, 1996, 2040, 2187, 2851, 4311, 4502 |
| Peak Set 10 | 10 | 100% | 99.90% | 99.96% | 1580, 1996, 2040, 2187, 2411, 2851, 3604, 4311, 4502, 4516 |
a
Results were obtained by calculating the average prediction accuracy of 500 test sets for each peak set.