Research on a Pediatric Common Disease Question Answering System Based on Knowledge Graph

Due to the characteristics of pediatric diseases such as their variety and complex etiology, the diagnosis of pediatric diseases often requires doctors to have rich professional knowledge and clinical experience. However, in reality, doctors in grassroots medical institutions and areas with scarce medical resources often lack diagnostic experience and find it difficult to make accurate judgments quickly. To address the above issues, this paper introduces a method for implementing a question answering system based on a knowledge graph of common pediatric diseases. The system uses a BERT-Bi-LSTM-CRF model for entity recognition and leverages the Neo4j graph database to store and visualize the resulting knowledge graph of common pediatric diseases. This paper utilizes pediatric disease question answering texts from the Chinese Medical Dialogue Dataset as the data source. The experimental outcomes indicate that the question answering system for common pediatric diseases, based on the method described, can improve the knowledge extraction from question answering texts, effectively converting unstructured community medical question answering texts into structured data, and aids in the auxiliary diagnosis of common pediatric diseases.