Skip to main content

Table 2 Comparison of algorithms for BOA training

From: Cascade of Boolean detector combinations

BOA

BOATS

IES

BAROC

BOR

68.3

68.6

68.2

BAND,N=1

75.7

76.5

73.5

¬B¬OR

76.3

73.2

73.2

BAND,N=6

81.1

77.3

73.3

\(\neg \mathrm {B}_{\neg \mathcal {P}}, N_{q}=1\)

81.3

73.4

73.7

\(\mathrm {B}_{\mathcal {P}}, N_{q}=1\)

80.0

81.8

80.5

\(\mathrm {B}_{\mathcal {P}}, N_{q}\) by BOATS

81.7

  
  1. Average test F1-score over sixfold cross-validation sets in context change detection task with BOA combinations (17) trained with different algorithms. The used operating point of the BOA is the one with highest F1-score on train data separately for each CV-fold