Construction of traditional Chinese medicine Knowledge Graph using Data Mining and Expert Knowledge

Abstract

Knowledge Graph (KG) is a powerful tool for Medical Decision Support(MDS). In this paper, we propose a novel method to construct a knowledge graph of traditional Chinese medicine (TCM), which combines data mining on limited but typical electronic medical records (EMRs) with expert knowledge. In particular, we leverage data mining to mine the medical regulations, and then convert them into medical knowledge with the help of experts, and finally build the KG accordingly. The goal of data mining is to exclude infrequent patterns which are caused by interference in EMRs. This method takes a different view from knowledge extraction. It avoids over-intervention of experts and subjectivity resulting from experts' direct intervention. This method can also be applied to other professional fields whose training data includes foundation knowledge and empirical information. The constructed KG has been approved by expert evaluation.