Commit
•
cd80e53
1
Parent(s):
b10be2f
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,45 +1,91 @@
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
- spacy
|
|
|
|
|
|
|
| 4 |
language:
|
| 5 |
- nl
|
| 6 |
license: cc-by-sa-4.0
|
| 7 |
model-index:
|
| 8 |
-
- name:
|
| 9 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
---
|
| 11 |
-
Package to classify spans for the presence and severity of left ventricular systolic dysfunction in Dutch echocardiogram reports.
|
| 12 |
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
-
|
| 26 |
|
| 27 |
<details>
|
| 28 |
|
| 29 |
-
<summary>View label scheme (
|
| 30 |
|
| 31 |
| Component | Labels |
|
| 32 |
| --- | --- |
|
| 33 |
-
| **`spancat`** | `lv_sys_func_normal`, `lv_sys_func_mild`, `lv_sys_func_moderate`, `lv_sys_func_severe
|
| 34 |
|
| 35 |
</details>
|
| 36 |
|
| 37 |
-
### Accuracy
|
| 38 |
|
| 39 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
| --- | --- |
|
| 41 |
-
|
|
| 42 |
-
|
|
| 43 |
-
|
|
| 44 |
-
|
|
| 45 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
- spacy
|
| 4 |
+
- arxiv:2408.06930
|
| 5 |
+
- medical
|
| 6 |
language:
|
| 7 |
- nl
|
| 8 |
license: cc-by-sa-4.0
|
| 9 |
model-index:
|
| 10 |
+
- name: Echocardiogram_SpanCategorizer_lv_syst_func
|
| 11 |
+
results:
|
| 12 |
+
- task:
|
| 13 |
+
type: token-classification
|
| 14 |
+
dataset:
|
| 15 |
+
type: test
|
| 16 |
+
name: "internal test set"
|
| 17 |
+
metrics:
|
| 18 |
+
- name: "Weighted f1"
|
| 19 |
+
type: f1
|
| 20 |
+
value: 0.770
|
| 21 |
+
verified: false
|
| 22 |
+
- name: "Weighted precision"
|
| 23 |
+
type: precision
|
| 24 |
+
value: 0.792
|
| 25 |
+
verified: false
|
| 26 |
+
- name: "Weighted recall"
|
| 27 |
+
type: recall
|
| 28 |
+
value: 0.749
|
| 29 |
+
verified: false
|
| 30 |
+
|
| 31 |
+
pipeline_tag: token-classification
|
| 32 |
+
metrics:
|
| 33 |
+
- f1
|
| 34 |
+
- precision
|
| 35 |
+
- recall
|
| 36 |
---
|
|
|
|
| 37 |
|
| 38 |
+
# Description
|
| 39 |
+
This model is a spaCy SpanCategorizer model trained from scratch on Dutch echocardiogram reports sourced from Electronic Health Records. The publication associated with the span classification task can be found at https://arxiv.org/abs/2408.06930. The config file for training the model can be found at https://github.com/umcu/echolabeler.
|
| 40 |
+
|
| 41 |
+
# Minimum working example
|
| 42 |
+
```python
|
| 43 |
+
!pip install https://huggingface.co/baukearends/Echocardiogram-SpanCategorizer-lv-syst-func/resolve/main/nl_Echocardiogram_SpanCategorizer_lv_syst_func-any-py3-none-any.whl
|
| 44 |
+
```
|
| 45 |
+
```python
|
| 46 |
+
import spacy
|
| 47 |
+
nlp = spacy.load("nl_Echocardiogram_SpanCategorizer_lv_syst_func")
|
| 48 |
+
```
|
| 49 |
+
```python
|
| 50 |
+
prediction = nlp("Op dit echo geen duidelijke WMA te zien, goede systolische L.V. functie, wel L.V.H., diastolische dysfunctie graad 1A tot 2. Geringe aortastenose en - matige -insufficientie. Geringe M.I.")
|
| 51 |
+
for span, score in zip(prediction.spans['sc'], prediction.spans['sc'].attrs['scores']):
|
| 52 |
+
print(f"Span: {span}, label: {span.label_}, score: {score[0]:.3f}")
|
| 53 |
+
```
|
| 54 |
|
| 55 |
+
# Label Scheme
|
| 56 |
|
| 57 |
<details>
|
| 58 |
|
| 59 |
+
<summary>View label scheme (4 labels for 1 components)</summary>
|
| 60 |
|
| 61 |
| Component | Labels |
|
| 62 |
| --- | --- |
|
| 63 |
+
| **`spancat`** | `lv_sys_func_normal`, `lv_sys_func_mild`, `lv_sys_func_moderate`, `lv_sys_func_severe` |
|
| 64 |
|
| 65 |
</details>
|
| 66 |
|
|
|
|
| 67 |
|
| 68 |
+
# Intended use
|
| 69 |
+
The model is developed for span classification on Dutch clinical text. Since it is a domain-specific model trained on medical data, it is meant to be used on medical NLP tasks for Dutch.
|
| 70 |
+
|
| 71 |
+
# Data
|
| 72 |
+
The model was trained on approximately 4,000 manually annotated echocardiogram reports from the University Medical Centre Utrecht. The training data was anonymized before starting the training procedure.
|
| 73 |
+
|
| 74 |
+
| Feature | Description |
|
| 75 |
| --- | --- |
|
| 76 |
+
| **Name** | `Echocardiogram_SpanCategorizer_lv_syst_func` |
|
| 77 |
+
| **Version** | `1.0.0` |
|
| 78 |
+
| **spaCy** | `>=3.7.4,<3.8.0` |
|
| 79 |
+
| **Default Pipeline** | `tok2vec`, `spancat` |
|
| 80 |
+
| **Components** | `tok2vec`, `spancat` |
|
| 81 |
+
| **License** | `cc-by-sa-4.0` |
|
| 82 |
+
| **Author** | [Bauke Arends]() |
|
| 83 |
+
|
| 84 |
+
# Contact
|
| 85 |
+
If you are having problems with this model please add an issue on our git: https://github.com/umcu/echolabeler/issues
|
| 86 |
+
|
| 87 |
+
# Usage
|
| 88 |
+
If you use the model in your work please use the following referral; https://doi.org/10.48550/arXiv.2408.06930
|
| 89 |
+
|
| 90 |
+
# References
|
| 91 |
+
Paper: Bauke Arends, Melle Vessies, Dirk van Osch, Arco Teske, Pim van der Harst, René van Es, Bram van Es (2024): Diagnosis extraction from unstructured Dutch echocardiogram reports using span- and document-level characteristic classification, Arxiv https://arxiv.org/abs/2408.06930
|