Table 1 Classification performance on registry screening data. Model performance is given as the probability of agreement [13] score (\(\Phi _s\)) with \(95 \%\) CI

GDL

\(0.35 \left[ 0.32, 0.43 \right]\)

\(0.087 \left[ 0.077, 0.094 \right]\)

\(0.15 \left[ 0.13, 0.17 \right]\)

\(0.47 \left[ 0.44, 0.51 \right]\)

1.1

MF

\(0.28 \left[ 0.22, 0.35 \right]\)

\(0.022 \left[ 0.00, 0.063 \right]\)

\(0.21 \left[ 0.19, 0.24 \right]\)

\(0.46 \left[ 0.33, 0.54 \right]\)

0.98

CMF

\(0.31 \left[ 0.23, 0.39 \right]\)

\(0.11 \left[ 0.063, 0.12 \right]\)

\(0.29 \left[ 0.27, 0.32 \right]\)

\(0.77 \left[ 0.72, 0.83 \right]\)

1.5

WCMF

\(0.31 \left[ 0.26, 0.35 \right]\)

\(0.25 \left[ 0.23, 0.27 \right]\)

\(0.27 \left[ 0.24, 0.31 \right]\)

\(0.78 \left[ 0.73, 0.87 \right]\)

1.6

SCMF

\(0.33 \left[ 0.27, 0.39 \right]\)

\(\mathbf {0.59 \left[ 0.57, 0.62 \right] }\)

\(\mathbf {0.35 \left[ 0.32, 0.37 \right] }\)

\(0.63 \left[ 0.55, 0.71 \right]\)

1.9

SWCMF

\(\mathbf {0.36 \left[ 0.29, 0.41 \right] }\)

\(0.50 \left[ 0.47, 0.51 \right]\)

\(0.33 \left[ 0.24, 0.41 \right]\)

\(\mathbf {0.86 \left[ 0.80, 0.90 \right] }\)

\(\mathbf {2.1}\)