Skip to main content
. 2014 Nov 6;10(11):e1003889. doi: 10.1371/journal.pcbi.1003889

Table 2. PPV using different inputs features.

Features Inline graphic Inline graphic
16 Predictions 0.57 0.67
16 Predictions + Separation 0.59 0.68
16 Predictions + SS 0.57 0.70
16 Predictions + PSSM 0.59 0.70
16 Predictions + RSA 0.59 0.71
16 Predictions + all features 0.61 0.73

Impact of different feature combinations on PPV values for first and fifth layer of deep learning.