Skip to main content

Table 4 Multiple grammar classifiers

From: Probabilistic grammatical model for helix‐helix contact site classification

Class

WDG150NR

PDBTM150NR

 

Combined

Multi

Combined

Multi

c1

0.56 (0.59)

0.66

0.47 (0.48)

0.50

c2

0.67 (0.68)

0.62

0.54 (0.53)

0.54

c3

0.64 (0.71)

0.65

0.60 (0.59)

0.57

c4

0.57 (0.70)

0.66

0.57 (0.55)

0.61

avg

0.60

0.65

0.52

0.54

  1. Average AUC ROC of the combined and multiple grammar classifiers in WDG150NR (4‐CV) and PDBTM150NR datasets. All results refer to grammars trained using WDG150NR in the 4‐CV experiment. Test performance of the single grammars, which were the best in 4‐fold CV in WDG150NR, is shown in parentheses. avg is an average weighted by relative numbers of samples in each class.
Back to article page

Algorithms for Molecular Biology

ISSN: 1748-7188