Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
metrics:
|
| 6 |
+
- f1
|
| 7 |
+
- accuracy
|
| 8 |
+
library_name: transformers
|
| 9 |
+
pipeline_tag: text-classification
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# lncrna-biocontext
|
| 13 |
+
This model is designed to determine whether a given abstract talks about an lncRNA in the context of disease or not.
|
| 14 |
+
|
| 15 |
+
The model has been trained on data from [lncBook-Wiki](https://ngdc.cncb.ac.cn/lncbook/) about papers
|
| 16 |
+
which have been curated by experts based on the biological context they discuss. We have collected the
|
| 17 |
+
abstracts for these papers and simplified the classification into disease/not disease. We then fine-tune a
|
| 18 |
+
[longformer](https://huggingface.co/allenai/longformer-base-4096) model to make a binary classification.
|
| 19 |
+
|
| 20 |
+
We achieve pretty good results:
|
| 21 |
+
|
| 22 |
+
| Metric | Score |
|
| 23 |
+
|-|- |
|
| 24 |
+
| Accuracy | 0.84 |
|
| 25 |
+
| F1 | 0.82 |
|
| 26 |
+
| ROC| 0.98 |
|
| 27 |
+
|
| 28 |
+
Though the test set is only 59 examples, with 22 discussing disease.
|
| 29 |
+
|
| 30 |
+
The next step will be to be able to classify both the specific disease (e.g. lung adenocarcinoma), and the non-disease
|
| 31 |
+
context (e.g. localisation) a paper discusses.
|
| 32 |
+
|
| 33 |
+
|