Representation Learning of EHR Data via Graph-Based Medical Entity Embedding
This work addresses the need for structured and actionable information in healthcare informatics by providing a method to represent key EHR entities, though it appears incremental as it builds on existing graph embedding techniques.
The paper tackled the problem of automatically learning representations of medical entities from EHR data, proposing ME2Vec, a graph-based embedding framework that improved disease diagnosis prediction over baselines using real-world clinical data.
Automatic representation learning of key entities in electronic health record (EHR) data is a critical step for healthcare informatics that turns heterogeneous medical records into structured and actionable information. Here we propose ME2Vec, an algorithmic framework for learning low-dimensional vectors of the most common entities in EHR: medical services, doctors, and patients. ME2Vec leverages diverse graph embedding techniques to cater for the unique characteristic of each medical entity. Using real-world clinical data, we demonstrate the efficacy of ME2Vec over competitive baselines on disease diagnosis prediction.