Table 2. Performance of individual feature-based models for the unbound sequence dataset, evaluated by LOOCV.
Window | #1 | #2 | #3 | #4 | #5 | #6 | #7 | #8 | #9 | Average |
5 | 0.572 | 0.522 | 0.572 | 0.575 | 0.639 | 0.571 | 0.546 | 0.600 | 0.617 | 0.579 |
7 | 0.592 | 0.544 | 0.585 | 0.592 | 0.632 | 0.575 | 0.522 | 0.609 | 0.624 | 0.586 |
9 | 0.603 | 0.543 | 0.581 | 0.585 | 0.635 | 0.575 | 0.531 | 0.616 | 0.627 | 0.588 |
11 | 0.606 | 0.556 | 0.601 | 0.597 | 0.633 | 0.579 | 0.541 | 0.611 | 0.626 | 0.595 |
13 | 0.606 | 0.558 | 0.584 | 0.583 | 0.625 | 0.572 | 0.543 | 0.595 | 0.621 | 0.588 |
15 | 0.604 | 0.520 | 0.584 | 0.581 | 0.626 | 0.577 | 0.554 | 0.586 | 0.626 | 0.584 |
Physicochemical propensities (#1), amino acid composition (#2), amino acid function group (#3), amino acid functional composition (#4), evolutionary profile (#5), sparse profile (#6), amino acid pair profile (#7), secondary structure (#8), relative accessible surface area (#9).