There is a newer version of the record available.

Published June 15, 2021 | Version v4
Journal article Open

A Comparative Dataset: Bridging COVID-19 and Other Diseases through Epistemonikos and CORD-19 Evidence

  • 1. National Center for Artificial Intelligence
  • 2. Pontificia Universidad Católica de Chile

Description

This is the dataset for COVID-19 document type screening. 

It is composed of: 

- Epistemonikos train dataset 

- CORD-19 test dataset adapted for Evidence Based Medicine domain

- XLNET model fine-tuned on Epistemonikos dataset. 

- BioBERT model fine-tuned on Epistemonikos dataset.

Scripts to run experiments can be found at: https://github.com/afcarvallo/covid_19_document_type_screening

Files

CORD19_full_labels.csv

Files (1.7 GB)

Name Size Download all
md5:dd7a8615c6977b17f80e97b777cae005
433.3 MB Download
md5:f6fc88612a8d371852606c77c7bd1aa9
36.2 MB Preview Download
md5:26747b89d80a66053149328442b8fe79
760.8 MB Download
md5:959830a44a0a2fc391e6a47aeaa986b9
466.9 MB Download