Table 7 Performance comparison of the three models
From: Logical definition-based identification of potential missing concepts in SNOMED CT
Metric | BART | T5 | PEGASUS |
|---|---|---|---|
ROUGE-1 | 72.66 | 72.67 | 72.83 |
ROUGE-2 | 50.79 | 50.15 | 51.06 |
ROUGE-L | 71.58 | 71.64 | 71.76 |
No. of correct prediction | 11,475/36,146 | 11,521/36,146 | 11,549/36,146 |