/FLAG

Source code and software artifacts for the paper "FLAG: Financial Long Document Regression via AMR-based GNN‘

Primary LanguagePythonMIT LicenseMIT

FLAG

Source code and software artifacts for the method Financial Long document classfication via AMR-based GNNs (FLAG), proposed in the paper "Semantic Graph Learning for Trend Prediction from Long Financial Documents‘

Model Weights and Data

The model weights for FLAG on the Earnings Call dataset are available here: DOI.

The csv files containing the Earnings Call dataset to be used directly by the Python scripts are available here: DOI.

The AMR files for FLAG from the Earnings Call dataset are available here: DOI