Model Card for incar-status-any

A Clinical Longformer-based model trained by the HAIL lab to predict incarceration status (past and present) in ED Notes.

Model Details

Model Description

• Developed by: Vimig Socrates
• Model type: Longformer
• Language(s) (NLP): English
• License: Apache License 2.0
• Finetuned from model: Clinical Lonformer

Uses

This model can be used to predict the incarceration status that a patient might have given most types of clinical ED notes.

Bias, Risks, and Limitations

This should not be used directly without supervision from a physician as predicting incarceration status incorrectly can have significant negative social and clinical impacts.

Training Details

Training Data

This model was trained on custom annotated data labeled for incarceration status from Yale-New Haven Health Hospital System ED Notes.

Training Procedure

Evaluation

TODO

Testing Data, Factors & Metrics

Results

TODO

]

Citation [optional]

Coming soon!

BibTeX:

{{ citation_bibtex | default("[More Information Needed]", true)}}

APA:

{{ citation_apa | default("[More Information Needed]", true)}}

Model Card Authors [optional]

Vimig Socrates

Model Card Contact

Vimig Socrates: vimig.socrates@yale.edu