Code Synonyms Do Matter: Multiple Synonyms Matching Network for Automatic ICD Coding

Anonymous

Abstract: Automatic ICD coding is defined as assigning disease codes to electronic medical records (EMRs).Existing methods apply label attention with code representations to match related text snippets for coding.Unlike these works that model the label with the code hierarchy or description, we argue that the code synonyms can provide more comprehensive knowledge based on the observation that the code expressions in EMRs vary from their descriptions in ICD. By aligning codes to concepts in UMLS, we collect synonyms of every code in ICD. Then, we propose a multiple synonyms matching network to leverage synonyms for better code representation learning, and finally help the code classification. Experiments on two settings of the MIMIC-III dataset show that our proposed method outperforms previous state-of-the-art methods.