Table 3.
Performance Comparison for Neural Network Learning. Labels are as described in the caption for Table 2.
Metric | CMS | Threshold | ANN built after n hours of data | ||||
---|---|---|---|---|---|---|---|
|
1 | 2 | 4 | 8 | |||
Sensitivity | 1.00 | 1.00 | 0.05 [0.0,.11] | 0.23 [0.0,.29] | 0.60 [.39,.66] | 0.80 [.61,.90] | 0.96 [.81,.98] |
Specificity | 0.99 | 0.88 | 0.99 [.97,1.0] | 0.97 [.84,.99] | 0.98 [.94,.99] | 0.99 [.95,.99] | 0.99 [.96,.99] |
PPV | 0.82 | 0.70 | 0.70 [0.0, .74] | 0.16 [0.0,.24] | 0.37 [.02,.72] | 0.71 [.49,.78] | 0.79 [.63,.80] |
Accuracy | 0.99 | 0.96 | 0.96 [.92,.96] | 0.95 [.90,.96] | 0.96 [.90,.99] | 0.97 [.92,.99] | 0.99 [.95,1.0] |