Assertion Status Detection on i2b2 VA clinical challenge 2010 (test)
94.4Accuracy (Absent)Spark NLP
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Spark NLPannotator=AssertionDLModel, architecture=Bi-LSTM, learning rate=0.0012, dropout=0.05, batch size=64, maximum sentence length=250, training epochs=20, window size=9 left, 15 right2020.12 | 94.4 | 90.4 | 44.1 | 86.2 | 68 | 95.3 | 93.9 | — | |
| Latest Best2020.12 | 93.7 | 86.9 | 42.2 | 89 | 63 | 95.7 | 93.4 | — | |
| Baseline_BlueBERTBackbone=BlueBERT2026.03 | — | — | — | — | — | — | — | 94.3 | |
| i2b2/VA Best2026.03 | — | — | — | — | — | — | — | 93.6 |