---
license: cc-by-4.0
---

# Pat2Vec
Fro a description of the framework and model, see our publication: <https://preprints.jmir.org/preprint/40755/>

It is trained using the amazing gensim package version 4 and parameters were optimized with Bayesian optimization (using another amazing package, optuna).

Unfortunately, this gensim model cannot be loaded directly. You have to clone the repository or download all the files and run the following to use it.

## quick start to use the model in Python:
```
from gensim.models.doc2vec import Doc2Vec

pat2vec_model = Doc2Vec.load('pat2vec_dim10.model')

pat2vec_model.infer_vector(["M54.1", "J06.9", "I10.90", "R51"])
```