Table 4 Precision, Recall, and F1-scores for each label on the held-out test data
From: Synthetic data for annotation and extraction of family history information from clinical text
Label | Precision | Recall | F1-score | Nr. of instances |
|---|---|---|---|---|
AGE | 0.797 | 0.505 | 0.618 | 93 |
AMOUNT | 0.618 | 0.778 | 0.689 | 81 |
CONDITION | 0.651 | 0.651 | 0.651 | 261 |
EVENT | 0.511 | 0.630 | 0.564 | 73 |
FAMILY | 0.706 | 0.859 | 0.775 | 249 |
INDEX | 0.000 | 0.000 | 0.000 | 7 |
NEG | 0.421 | 0.571 | 0.485 | 14 |
O | 0.908 | 0.872 | 0.890 | 2066 |
SELF | 0.929 | 0.730 | 0.818 | 126 |
TEMPORAL | 0.425 | 0.761 | 0.545 | 67 |
Weighted Average (SVM) | 0.835 | 0.821 | 0.824 | 3037 |
Dictionary baseline | 0.770 | 0.607 | 0.647 | 3037 |
SVM (excluding “O” label) | 0.678 | 0.712 | 0.684 | 971 |
Dictionary baseline (excluding “O” label) | 0.543 | 0.720 | 0.581 | 971 |