Named Entity Recognition for COVID-19 Bio Entities

The dataset was taken from https://github.com/davidcampos/covid19-corpus

Dataset

The dataset was then split into several datasets each one representing one entity. Namely, Disorder, Species, Chemical or Drug, Gene and Protein, Enzyme, Anatomy, Biological Process, Molecular Function, Cellular Component, Pathway and microRNA. Moreover, another dataset is also created with all those aforementioned that are non-overlapping in nature.

Dataset Formats

The datasets are available in two formats IOB and Spacy's JSONL format.

IOB : https://github.com/tsantosh7/COVID-19-Named-Entity-Recognition/tree/master/Datasets/BIO

SpaCy JSONL: https://github.com/tsantosh7/COVID-19-Named-Entity-Recognition/tree/master/Datasets/SpaCy

Tutorials

1. Convert_to_JSON_IOB.ipynb : Coverts David Campos's BRAT annotations format to IOB and SpaCy's JSONL format

2. Train_NER_Model.ipynb : Trains SpaCy on the non-overlapping set for Named Entity Recognition (NER)

Performance

Chemicals or Drugs
precision-CHED:	0.90
recall-CHED:	0.93
f1-measure-CHED:	0.91
Cellular Component
precision-COMP:	0.90
recall-COMP:	0.95
f1-measure-COMP:	0.92
Molecular Function
precision-FUNC:	0.89
recall-FUNC:	0.94
f1-measure-FUNC:	0.92
Anatomy
precision-ANAT:	0.92
recall-ANAT:	0.93
f1-measure-ANAT:	0.92
Pathway
precision-PATH:	0.89
recall-PATH:	0.92
f1-measure-PATH:	0.90
Disorder
precision-DISO:	0.90
recall-DISO:	0.94
f1-measure-DISO:	0.92
Enzyme
precision-ENZY:	0.87
recall-ENZY:	0.88
f1-measure-ENZY:	0.88
Gene or Protein
precision-PRGE:	0.76
recall-PRGE:	0.77
f1-measure-PRGE:	0.77
Biological Process
precision-PROC:	0.92
recall-PROC:	0.94
f1-measure-PROC:	0.93
Species
precision-SPEC:	0.92
recall-SPEC:	0.95
f1-measure-SPEC:	0.94
O-tag
precision-untyped:	0.91
recall-untyped:	0.94
f1-measure-untyped:	0.92
Combined Overall Result
precision-overall:	0.90
recall-overall:	0.92
f1-measure-overall:	0.91

Using the model

download SpaCy model:

https://drive.google.com/open?id=1QoIeOAUZqaFLGrS-gJH1aAlE1_bn22ZI

install the model:

pip install en_covido-0.0.0.tar.gz 

load the model:

import spacy
import en_covido

perform NER

nlp = en_covido.load()

doc = nlp("The coronavirus pandemic has become a growing public health concern worldwide, and there are insufficient epidemiological data on critical illness. We sought to investigate the clinical course and features of critical patients with Corona Virus Disease 2019 (COVID-19).")

for ent in doc2.ents:
    print(ent.label_, ent.text)

check results

SPEC coronavirus
DISO critical illness
DISO Corona Virus Disease 2019
DISO COVID-19

Contributing

Pull requests are welcome.

Please make sure to update tests as appropriate.

Contact

If you would like to know more or contribute, please send an e-mail or submit a pull request.

License

MIT