An Approach and Implementation for Knowledge Graph Construction and QA System

ABSTRACT

In this paper, we investigate and propose a knowledge graph-based method and implementation of the question-and-answer (QA) system for COVID-19 cases imported from abroad. It mainly analyzes and organizes the knowledge graph construction methods based on knowledge acquisition and visualization. In addition, this paper implements the knowledge graph-based QA system by training term frequency-inverse document frequency (TF-IDF) model and Bidirectional Long Short-Term Memory + Conditional Random Field (Bi-LSTM+CRF) model as well as Cypher query statements using the graph database Neo4j. Finally, the visual intelligent interface of the QA system is designed to meet user requirements and realize the function of accurate QA. © 2021 IEEE.