Table 3.
Micro-averaged results on the development set (269 narratives).
Token level matching | Text strict matching | HIPAA strict matching | |||||||
---|---|---|---|---|---|---|---|---|---|
P% | R% | F% | P% | R% | F% | P% | R% | F% | |
Submission 1 | 97.56 | 88.71 | 92.92 | 94.26 | 84.03 | 88.85 | 96.63 | 84.06 | 89.91 |
Submission 2 | 96.8 | 93.99 | 95.37 | 92.56 | 89.11 | 90.8 | 95.21 | 91.34 | 93.24 |
Submission 3 | 97.63 | 93.11 | 95.31 | 93.73 | 88.68 | 91.13 | 95.65 | 90.94 | 93.23 |