Download the CORD-19 dataset from here. The data should be in ./cord19
.
Install dependencies:
python -m venv venv
source venv/bin/activate
pip install --upgrade pip setuptools
pip install -r requirements.txt
Start neo4j server (only needed if running analysis and plotting neo4j graph):
bash neo4j-community-4.1.2/bin/neo4j start
Run jupyter and open notebook:
jupyter notebook
To reproduce the results, follow the steps below:
- Fine-tune SciBERT on SciCite (citation_intent_classification.ipynb)
- Process the CORD-19 data (process_data.ipynb)
- Run the anaylsis (analysis.ipynb)