Skip to main content
. 2008 Nov 12;25(1):30–35. doi: 10.1093/bioinformatics/btn583

Table 1.

The prediction performance of the RF model based on various features. The prediction system was evaluated by nested cross-validation and the threshold for sample classification is 1/6

Features ACC±SD(%) SE±SD(%) PR±SD(%) SP±SD(%) MCC
A 88.93±0.19 76.29±1.40 63.39±0.79 91.46±0.22 0.633
A+B 89.69±0.41 76.58±1.94 66.51±1.28 92.31±0.35 0.652
A+C 90.46±0.44 75.64±0.65 69.63±2.36 93.41±0.59 0.668
A+B+C 91.41±0.49 76.57±1.69 73.16±2.78 94.38±0.76 0.70

A: PSSMs; B: SS; C: OBVs of amino acids.