phenobert 1.0.5

A novel tool for human clinical disease phenotype recognizing with deep learning.

PhenoBERT

A novel tool for human clinical disease phenotype recognizing with deep learning.

What is PhenoBERT?

PhenoBERT is a method that uses advanced deep learning methods (i.e. convolutional neural networks and BERT) to identify clinical disease phenotypes from free text. Currently, only English text is supported. Compared with other methods in the expert-annotated test set, PhenoBERT has reached SOTA effect.

In GSC+(Lobo et al., 2017) dataset:

Method	Precision	Recall	F1-score	Set Similarity
NCBO Annotator	75.22	49.05	59.38	72.22
NeuralCR	72.51	58.71	64.88	78.93
Clinphen	55.26	38.13	45.12	55.86
MetaMapLite	67.21	44.25	53.37	67.07
PhenoPro-NLP	14.47	53.32	22.76	34.64
PhenoBERT	77.97	60.70	68.26	82.41

How to install PhenoBERT

pip

PhenoBERT supports Python 3.6 or later. We recommend that you install PhenoBERT via pip, the Python package manager. To install, simply run:

pip install phenobert

This should also help resolve all of the dependencies of PhenoBERT, for instance PyTorch 1.3.0 or above.

From Source

Pretrained embeddings and models

We have trained