Research And Implementation Of The Question Answering System Of The Kingdoms In Eastern Zhou Based On Knowledge Graph

Chinese civilization has a history of 5000 years and a long history.From the Xia,Shang,Western Zhou,to the Eastern Zhou and other kingdoms,to the Qin,Han,Wei,Jin,Tang,Song,Yuan,Ming,and Qing dynasties,Chinese civilization has experienced prosperity and decline,as well as countless years of ups and downs.The Eastern Zhou Dynasty was an important period in the history of the pre Qin period and one of the key periods for the development of Chinese civilization.During this period,significant progress was made in China’s social economy,culture,art,and philosophy of thought.Considering the gap in the knowledge graph during the Eastern Zhou Dynasty,this article constructs a knowledge graph of the Eastern Zhou Dynasty and implements a question and answer system for the Eastern Zhou Dynasty based on the completion of the knowledge graph.The main research content of this article is as follows:(1)Build a knowledge map of the Eastern Zhou states.The experimental data were obtained from two parts,one of which used the crawler technology to crawl the relevant information of the countries of the Eastern Zhou Dynasty from Baidu Baike,Wikipedia and other encyclopedia websites,the other part extracted the text of the book Annals of the Countries of the Eastern Zhou Dynasty,simplified the text of the whole book,and then extracted data information from the simplified text.Clean all the data and extract knowledge extraction to form a knowledge triplet.Perform knowledge fusion on the data,use the fused data to construct a knowledge graph of the Eastern Zhou countries,and store the constructed knowledge graph in the graph database.(2)Research on named entity recognition algorithms.Collect datasets from various countries in the Eastern Zhou Dynasty,clean the data,and then perform Chinese word segmentation and data annotation.By studying named entity recognition algorithms,it was ultimately determined to conduct experiments primarily with the BERT-BILSTM-CRF model,introducing attention mechanisms into the BILSTM model,changing high weights,and focusing on entity recognition.Annotate data,train the model,adjust model parameters,and then process the data using the trained model to identify entities of the Eastern Zhou Dynasty.During the experiment,different annotation mechanisms were used for data processing,as well as different named entity recognition models to complete comparative experiments.The experimental results demonstrate that the BERT-BILSTM-CRF model with attention mechanism achieves better performance in entity recognition tasks compared to other models.(3)Build a complete Q&A system.On the basis of the above research,by analyzing common problems,a question template is established,and then the question answering system of the Eastern Zhou countries is constructed through template matching and semantic analysis methods.Specifically,Chinese word segmentation is performed on user questions,questions raised by users are analyzed,problems are classified,and problem query statement templates are constructed.Then,the trained named entity recognition model is used to identify the entities in the question,and syntactic dependency analysis is performed to extract the relationships between entities,forming a question triplet.Entity mapping is performed with the constructed knowledge graph,matching the problem template,and forming a complete query statement.Then,knowledge retrieval is performed to obtain the constructed knowledge graph of the Eastern Zhou countries.Encapsulate the obtained results through visualization frameworks and tools,request access in the browser,and achieve automatic Q&A and query knowledge graph visualization functions of the system.Based on the above work,this article first constructs a knowledge graph of the Eastern Zhou countries,uses deep learning technology to enhance the semantic understanding of user questions in the question answering system,and ultimately implements a knowledge service platform for the Eastern Zhou countries,including a knowledge graph and a question answering system.This makes it convenient for users to learn the knowledge of the Eastern Zhou countries,and has also made certain contributions to research in related fields.