README.md

---
language: nl
license: mit
pipeline_tag: text-classification
inference: false
---

# A-PROOF ICF-domains Classification

## Description

A fine-tuned multi-label classification model that detects 17 [WHO-ICF](https://www.who.int/standards/classifications/international-classification-of-functioning-disability-and-health) domains in clinical text in Dutch. 
The model is based on a pre-trained Dutch medical language model ([link to be added]()), a RoBERTa model, trained from scratch on clinical notes of the Amsterdam UMC.

## ICF domains
The model can detect 17 categories, which were chosen due to their relevance to recovery from COVID-19:


ICF code | Domain | name in repo
---|---|---
b1300 | Energy level | ENR
b140 | Attention functions | ATT
b152 | Emotional functions | STM
b440 | Respiration functions | ADM
b455 | Exercise tolerance functions | INS
b530 | Weight maintenance functions | MBW
d450 | Walking | FAC
d550 | Eating | ETN
d840-d859 | Work and employment | BER
B280 | Sensations of pain | SOP
B134 | Sleep functions | SLP
D760 | Family relationships | FML
B164 | Higher-level cognitive functions | HLC
D465 | Moving around using equipment | MAE
D410 | Changing basic body position | CBP
B230 | Hearing functions | HRN
D240 | Handling stress and other psychological demands | HSP


## Intended uses and limitations
- The model was fine-tuned (trained, validated and tested) on medical records from the Amsterdam UMC (the two academic medical centers of Amsterdam). It might perform differently on text from a different hospital or text from non-hospital sources (e.g. GP records).
- The model was fine-tuned with the [Simple Transformers](https://simpletransformers.ai/) library. This library is based on Transformers but the model cannot be used directly with Transformers `pipeline` and classes; doing so would generate incorrect outputs. For this reason, the API on this page is disabled.

## How to use
To generate predictions with the model, use the [Simple Transformers](https://simpletransformers.ai/) library:
```
from simpletransformers.classification import MultiLabelClassificationModel

model = MultiLabelClassificationModel(
    'roberta',
    'CLTL/icf-domains',
    use_cuda=False,
)

example = 'Nu sinds 5-6 dagen progressieve benauwdheidsklachten (bij korte stukken lopen al kortademig), terwijl dit eerder niet zo was.'
predictions, raw_outputs = model.predict([example])
```
The predictions look like this:
```
[[0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]
```
The indices of the multi-label stand for:
```
[ENR-B1300, ATT-B140, STM-B152, ADM-B440, INS-B455, MBW-B530, FAC-D540, ETN-D550, BER-D840-D859, SOP-B280, SLP-B134, FML-D760, HLC-B164, MAE-D465, CBP-D410, HRN-B230, HSP-D240]
```
In other words, the above prediction corresponds to assigning the labels ADM, FAC and INS to the example sentence.

The raw outputs look like this:
```
[[0.51907885 0.00268032 0.0030862  0.03066113 0.00616694 0.64720929
  0.67348498 0.0118863  0.0046311 ]]
```
For this model, the threshold at which the prediction for a label flips from 0 to 1 is **0.5**.

## Training data
- The training data consists of clinical notes from medical records (in Dutch) of the Amsterdam UMC. Due to privacy constraints, the data cannot be released.
- The annotation guidelines used for the project can be found [here](https://github.com/cltl/a-proof-zonmw/tree/main/resources/annotation_guidelines).

## Training procedure
The default training parameters of Simple Transformers were used, including:
- Optimizer: AdamW
- Learning rate: 4e-5
- Num train epochs: 1
- Train batch size: 8
- Threshold: 0.5


## Authors and references
### Authors
Jenia Kim, Piek Vossen

### References

Kim, Jenia, Stella Verkijk, Edwin Geleijn, Marieke van der Leeden, Carel Meskers, Caroline Meskers, Sabina van der Veen, Piek Vossen, and Guy Widdershoven. "Modeling Dutch medical texts for detecting functional categories and levels of COVID-19 patients." In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pp. 4577-4585. 2022

@inproceedings{kim2022modeling,
  title={Modeling Dutch medical texts for detecting functional categories and levels of COVID-19 patients},
  author={Kim, Jenia and Verkijk, Stella and Geleijn, Edwin and van der Leeden, Marieke and Meskers, Carel and Meskers, Caroline and van der Veen, Sabina and Vossen, Piek and Widdershoven, Guy},
  booktitle={Proceedings of the Thirteenth Language Resources and Evaluation Conference},
  pages={4577--4585},
  year={2022}
}