LuckyLook is a tool designed for recommending PubMed scientific journals for researchers looking to publish their work. It uses a Transformer model with additional graph-level information to provide a better recommendation.

1. Download the dataset from Zenodo: https://doi.org/10.5281/zenodo.8386011
2. Create a Dataset/ folder in the LuckyLook directory.
3. Move the downloaded dataset into the Dataset/ folder.

Configuration JSON files, which include parameters and settings for model training, are stored in the configs/ directory.

To train the model, run:

python train.py --config configs/<config_file>.json

For testing, specify the checkpoint directory:

python test.py --resume path/to/checkpoint

streamlit run LuckyLook.py

• DGC-BERT: https://github.com/ECNU-Text-Computing/DGC-BERT