A Question Answering System of Ethnic Minorities Based on Knowledge Graph

Abstract

In recent years, Question Answering System has become a main focus of human machine interaction. Using the question answering system for information retrieval is convenient and efficient. Traditional question answering systems mostly use template matching. The question and answer data sets usually rely on manual design. The question and answer system implemented by this method has a quick query response and can answer relatively complex questions. But manually defining templates and rules is time-consuming and laborious. Therefore, this paper introduces the knowledge graph in the process of constructing the question answering system. The proposed method combines knowledge graph with question answering systems to realize an intelligent question answering system in the field of ethnic minorities, so as to spread ethnic minority knowledge and promote ethnic minority culture. The construction of a question answering system is mainly divided into two modules: question analysis and answer generation. The question analysis module includes named entity recognition, similarity calculation and question classification. The answer generation module includes entity mapping and graph retrieval. The experimental results show that our approach by combined with natural language processing technology and domain knowledge graph can well accurately feedback the answers to user queries. And the answer accuracy of the question answering system can reach about 81%.