A Question-Answering System Based on a Diabetes Knowledge Graph and Its Construction Method

Abstract

This invention discloses a question-answering system and its construction method based on a diabetes knowledge graph, belonging to the field of artificial intelligence technology. The method includes: obtaining triples and constructing a diabetes knowledge graph through text acquisition, cleaning and structure enhancement, entity relation extraction, and model training; parsing user questions, identifying and linking entities, determining query intent, and converting natural language questions into structured queries; retrieving data from the graph based on the structured queries, and returning the results as natural language answers. This invention improves the standardization and accuracy of knowledge extraction through context-aware text enhancement, multi-feature fusion entity boundary localization, multi-view relation extraction models, and end-to-end collaborative training, realizing medical logical reasoning question-answering based on structured knowledge.

Images