Commit
•
5c0fa20
1
Parent(s):
c482ca6
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,25 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- chrisvoncsefalvay/vaers-outcomes
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
metrics:
|
| 8 |
+
- accuracy
|
| 9 |
+
- f1
|
| 10 |
+
- precision
|
| 11 |
+
- recall
|
| 12 |
+
library_name: transformers
|
| 13 |
+
pipeline_tag: text-classification
|
| 14 |
+
tags:
|
| 15 |
+
- medical
|
| 16 |
+
- pharmacovigilance
|
| 17 |
+
- vaccines
|
| 18 |
---
|
| 19 |
+
|
| 20 |
+
DAEDRA (Detecting Adverse Event Dispositions for Regulatory Affairs) is a pharmacovigilance language model intended to facilitate the rapid identification and extraction of high-consequence outcomes from passive pharmacovigilance reporting. It was trained on the VAERS data set, and focuses on three main outcomes:
|
| 21 |
+
|
| 22 |
+
* mortality (VAERS `DIED` flag);
|
| 23 |
+
* emergency room attendance (`ER_VISIT`); and
|
| 24 |
+
* hospitalisation (`HOSPITAL`).
|
| 25 |
+
|