A Framework for Automated Knowledge Graph Construction Towards Traditional Chinese Medicine

Abstract

Medical knowledge graph can potentially help knowledge discovery from clinical data, assisting clinical decision making and personalized treatment recommendation. This paper proposes a framework for automated medical knowledge graph construction based on semantic analysis. The framework consists of a number of modules including a medical ontology constructor, a knowledge element generator, a structured knowledge dataset generator, and a graph model constructor. We also present the implementation and application of the constructed knowledge graph with the framework for personalized treatment recommendation. Our experiment dataset contains 886 patient records with hypertension. The result shows that the application of the constructed knowledge graph achieves dramatic accuracy improvements, demonstrating the effectiveness of the framework in automated medical knowledge graph construction and application.