Research on the Construction of a Knowledge Graph for Miao Medicine Based on the BERT-BiLSTM-CRF Model

This study focuses on constructing a knowledge graph for Miao medicine to promote the understanding of traditional Miao cultural practices, enhance public health capabilities, and preserve this intangible cultural heritage. The BERT-BiLSTM-CRF model, combined with the manual review, was utilized for entity recognition, while SBERT technology was used for entity alignment. As a result, a comprehensive knowledge graph of Miao medicine was constructed, including 11,111 nodes and 143,751 relationships. This knowledge graph systematically organizes the complex knowledge system of Miao medicine, including diseases, symptoms, drugs, and their interrelations. It supports efficient knowledge querying, reasoning, and discovery, contributing valuable insights into applying knowledge graph technologies in traditional medicine. Moreover, this research offers a framework that can be adapted for constructing knowledge graphs in other conventional medical systems.