DKEC: Domain Knowledge Enhanced Multi-Label Classification for Diagnosis Prediction
This addresses the challenge of few-shot classification in medical domains like emergency services and electronic health records, offering an incremental improvement by integrating external knowledge.
The paper tackled the long-tail label distribution problem in multi-label text classification for medical diagnosis prediction by incorporating external knowledge from medical guidelines, resulting in DKEC outperforming state-of-the-art models, especially for few-shot classes, and enabling smaller language models to match large ones.
Multi-label text classification (MLTC) tasks in the medical domain often face the long-tail label distribution problem. Prior works have explored hierarchical label structures to find relevant information for few-shot classes, but mostly neglected to incorporate external knowledge from medical guidelines. This paper presents DKEC, Domain Knowledge Enhanced Classification for diagnosis prediction with two innovations: (1) automated construction of heterogeneous knowledge graphs from external sources to capture semantic relations among diverse medical entities, (2) incorporating the heterogeneous knowledge graphs in few-shot classification using a label-wise attention mechanism. We construct DKEC using three online medical knowledge sources and evaluate it on a real-world Emergency Medical Services (EMS) dataset and a public electronic health record (EHR) dataset. Results show that DKEC outperforms the state-of-the-art label-wise attention networks and transformer models of different sizes, particularly for the few-shot classes. More importantly, it helps the smaller language models achieve comparable performance to large language models.