Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# BERT MCN-Model using SMM4H 2017 (subtask 3) data
|
| 2 |
+
|
| 3 |
+
The model was trained using [clagator/biobert_v1.1_pubmed_nli_sts](https://huggingface.co/clagator/biobert_v1.1_pubmed_nli_sts) as a base and the smm4h dataset from 2017 from subtask 3.
|
| 4 |
+
|
| 5 |
+
## Dataset
|
| 6 |
+
|
| 7 |
+
See [here](https://github.com/olastor/medical-concept-normalization/tree/main/data/smm4h) for the scripts and datasets.
|
| 8 |
+
|
| 9 |
+
**Attribution**
|
| 10 |
+
|
| 11 |
+
Sarker, Abeed (2018), “Data and systems for medication-related text classification and concept normalization from Twitter: Insights from the Social Media Mining for Health (SMM4H)-2017 shared task”, Mendeley Data, V2, doi: 10.17632/rxwfb3tysd.2
|
| 12 |
+
|
| 13 |
+
### Test Results
|
| 14 |
+
|
| 15 |
+
- Acc: 89.44
|
| 16 |
+
- Acc@2: 91.84
|
| 17 |
+
- Acc@3: 93.20
|
| 18 |
+
- Acc@5: 94.32
|
| 19 |
+
- Acc@10: 95.04
|
| 20 |
+
|
| 21 |
+
Acc@N denotes the accuracy taking the top N predictions of the model into account, not just the first one.
|