Fine-tuning DRAGON Longformer with tuned hyperparameters on 500 synthetic clinical cases reportedly raises accuracy from 72.0% to 85.2%, but the reported metrics are internally inconsistent.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Enhancing Clinical Text Classification via Fine-Tuned DRAGON Longformer Models
Fine-tuning DRAGON Longformer with tuned hyperparameters on 500 synthetic clinical cases reportedly raises accuracy from 72.0% to 85.2%, but the reported metrics are internally inconsistent.