Chinese Document-Level Emergency Event Extraction Dataset and Corresponding Methods

Structured extraction of emergency event information can effectively enhance the ability to respond events. This article focuses on extracting Chinese document-level events, which entails addressing two key issues in this field: related datasets and problem role overlapping between candidate entities, has been overlooked existing DEE (document-level extraction) studies that predominantly employed sequence annotation for entity extraction. To tackle these challenges, we constructed a dataset (CDEEE) provides annotations argument scattering, multiple overlapping. Additionally, model named RODEE is proposed address tasks. employs independent modules represent head tail positions utilizes multiplication attention mechanism interact two, generating scoring matrix. Subsequently, role-overlapping entities are predicted facilitate completion Experiments were conducted our manually annotated dataset, CDEEE, results show solves among resulting improved performance model, with an F1 value 77.7%.