Research on Domain Knowledge Graph Construction and Complex Question Answering Approach
LI Huayu
WANG Jiakun
YAN Yang
LI Jiarui
Abstract：The existing mature Chinese knowledge graph question answering methods can only answer some simple questions well,and it is difficult to deal with complex questions such as aggregation,comparison,multi hop and counting.Taking the field of party history as an example,this paper uses semi-automatic technology to construct the knowledge graph of party history.Then,a multi-task question answering model combining entity recognition,path prediction,question type recognition and answer entity type prediction is proposed to process complex questions.Finally,the validity of the model is proved by experimental analysis and question answering test.As a knowledge graph of party history and a complex Chinese question answering,this paper makes a pre-liminary beneficial exploration.
Keywords：knowledge graphnatural language processingmulti-task question answering modelcomplex question answer-ing approach
Publication Date：2025-02-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：7( 358-363,383 )