From: Confidence-based laboratory test reduction recommendation algorithm
Benchmark | Target Coverage | Model Coverage | Normality | Stability | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Prev | AUC | Acc | Prec | AUPRC | Prev | AUC | Acc | Prec | AUPRC | |||
Our Model | 0.85 | 83.80% | 15.25% | 95.89% | 93.17% | 74.38% | 80.05% | 97.40% | 95.94% | 97.41% | 97.46% | 99.89% |
Without Time Embedding | 0.85 | 85.28% | 15.64% | 95.61% | 92.85% | 74.63% | 79.53% | 97.99% | 94.92% | 98.01% | 98.01% | 99.89% |
Without Attention | 0.85 | 84.75% | 15.99% | 92.68% | 90.11% | 85.53% | 72.99% | 97.95% | 95.44% | 97.96% | 97.99% | 99.90% |
Without Confidence Selection | - | - | 20.20% | 90.91% | 85.19% | 59.28% | 63.92% | 90.90% | 94.80% | 93.48% | 95.68% | 99.46% |
Vanilla LSTM | - | - | 20.20% | 88.40% | 82.64% | 58.10% | 61.04% | 90.90% | 94.80% | 92.91% | 94.65% | 99.47% |