Research On The Key Technology Of Constructing Knowledge Graph Of Red Revolution Figures Based On Multi-source Data Fusion

Red culture has a unique function of casting the soul and nurturing people in the process of nurturing people for the Party and cultivating talents for the country.During the revolutionary war period,behind every red revolutionary figure,there is a singable story.The temperament and spirit in them are worth learning and inheriting,especially for young people with important educational significance,which can guide and encourage them to establish a correct world view,outlook on life and values.With the continuous development of new technologies such as big data and knowledge graph,knowledge graph has become an effective technology for knowledge representation,organization,and management,providing useful technical support for the management of red cultural knowledge.However,red cultural resources are characterized by a wide range of sources and fragmentation,and it is extremely challenging to obtain these data resources,and to extract,verify and organize knowledge from them.In this thesis,the heroic figures of the Red Revolution are taken as the research object;the characteristics of the red heroic figures are analyzed after acquiring the data,the knowledge graph of the red revolutionary figures is constructed using natural language related technology,and it is visualized to show the vein relationship of the heroic figures in the form of a diagram.The specific research contents are as follows.Firstly,using red websites,red databases and Baidu encyclopedia as data sources,we adopt crawler technology,design a data collection system,analyze the commonality and differences of different web page structures,using different data crawling strategies,and combining with Scrapy framework to obtain semi-structured and unstructured data such as text and video,using red books as reference,clean and process the final data,and generating a multi-source corpus of red revolutionary heroes and heroines.Secondly,the multi-source red revolutionary characters corpus is used as the object to classify and elaborate the relationship and content in the characters respectively,and the knowledge graph of red revolutionary characters is designed and constructed.Nodes in the graph are extracted using bert,while edges and information derive from semi-structured websites,fror which relationships and attributes are extracted.The acquisition of the graph triad is completed.The knowledge graph of red revolutionary figures is stored and displayed using the graph database Neo4 j.Finally,based on the constructed knowledge graph and the storage of graph database,a good entity and relationship dictionary are constructed;the question answer system of red revolutionary figures is designed and implemented based on three steps,namely,question analysis,question resolution as well as question answer query.