Research on Large-Scale Knowledge Base Management Frameworks for Open-Domain Question Answering Systems

In recent years, there are increasingly question answering systems based on large-scale knowledge bases that can answer natural questions. In this paper, we analyze the performance and efficiency of different knowledge base management frameworks when retrieving information from large-scale knowledge bases. The data model is built in the structure of a directed graph with vertices denoted the entities and edges denoted their relationships. With RDF (Resource Description Framework) model, Neo4j and Apache Jena built Graph Database Platform and Triple Store respectively to present the meaning networks. We analyzed, measured, and discussed how they store data from a knowledge base in the industry. We briefly showed the strengths and limits of each tool via experiments. Based on particular aims, researchers can choose the appropriate database management framework for their applications in large-scale open domain question answering systems.