Table 2.
Experiment | Predictor kernel | AUC score |
Time s |
Confidence intervals |
---|---|---|---|---|
I–PKMR–Non-Sequence | k exp | 0.660 | 300 | [0.655, 0.665] |
k loc | 0.503 | 240 | [0.499, 0.507] | |
k phy | 0.775 | 240 | [0.771, 0.779] | |
k exp + k loc + k phy | 0.755 | 350 | [0.752, 0.759] | |
w 1 k exp + w 2 k loc + w 3 k phy | 0.799 | 420 | [0.791, 0.807] | |
II–PKMR–Sequence | k pfam | 0.797 | 450 | [0.793, 0.801] |
k motif | 0.782 | 430 | [0.778, 0.786] | |
k mer | 0.725 | 420 | [0.720, 0.731] | |
k pfam + k motif + k mer | 0.817 | 480 | [0.811, 0.823] | |
w 4 k pfam + w 5 k motif + w 6 k mer | 0.821 | 530 | [0.818, 0.824] | |
III–PKMR–Combined (sequence and non-sequence) |
k phy + k pfam | 0.812 | 470 | [0.809, 0.816] |
k exp + k loc + k phy + k pfam + k motif + k mer | 0.831 | 610 | [0.828, 0.834] | |
w 1 k exp + w 2 k loc + w 3 k phy + w 4 k pfam + w 5 k motif + w 6 k mer | 0.840 | 720 | [0.831, 0.849] | |
IV–pSVM–Non-Sequence | k exp | 0.791 | 9020 | [0.786, 0.796] |
k loc | 0.696 | 7800 | [0.692, 0.700] | |
k phy | 0.802 | 7980 | [0.797, 0.807] | |
k exp + k loc + k phy | 0.818 | 10,100 | [0.812, 0.824] | |
w 1 k exp + w 2 k loc + w 3 k phy | 0.877 | 10,121 | [0.871, 0.883] | |
V–pSVM–Sequence | k pfam | 0.887 | 12,060 | [0.879, 0.895] |
k motif | 0.868 | 12,000 | [0.859, 0.877] | |
k mer | 0.840 | 11,760 | [0.836, 0.844] | |
k pfam + k motif + k mer | 0.898 | 12,220 | [0.891, 0.905] | |
w 4 k pfam + w 5 k motif + w 6 k mer | 0.910 | 12,800 | [0.901, 0.919] | |
VI–pSVM–Combined (Sequence and non-sequence) |
k phy + k pfam | 0.890 | 12,100 | [0.882, 0.898] |
k exp + k loc + k phy + k pfam + k motif + k mer | 0.939 | 13,420 | [0.935, 0.944] | |
w 1 k exp + w 2 k loc + w 3 k phy + w4 k pfam + w5 k motif + w6 k mer | 0.940 | 14,010 | [0.934, 0.946] |
These are AUC score (area under the ROC curve as accuracy), time s (Execution times in seconds), and confidence intervals