Design And Implementation Of Text Visualization System Based On Entity And Relationship Recognition Technology

With the rapid development of Internet technology and the emergence of massive information,quickly and accurately obtaining useful infor-mation is one of the difficulties in information extraction and natural language processing research.Text visualization technology aims to deliver text information more effectively through the combination of graphics and text.Information graphics are the main form of presentation in text visualization,which has the characteristics of content independence,and becomes a common carrier of information.In recent years,with the development of natural language understanding and artificial intelligence technology,automatic generation of information graphics based on text semantic understanding has become a new research direction,and has ganined widespread attention in academia and industry.This paper focuses on visualizing text containing statistical data,converting the problem of extracting statistical information into the problem of entity and relationship recognition,describing the statistical relationship between different entities through knowledge triples and trying to obtain an accurate description of statistical data in a graphical way to reduce the redundancy of text information and improve the efficiency of information acquisition.Existing researches often ignore the overlapping of entities and the existence of semantic inclusion relationships between different entities when identifying entities and relationships.At the same time,the existing text infographic tools have a low degree of automation and personalization,which makes it difficult to meet diverse visualization needs.In response to the above problems,this paper proposes a text visualization technology based on entity and relationship recognition.In particular,a judgment based on the semantic relationship between entities is proposed to identify the statistical inclusion relationship between entities,which effectively solve the problem of entity overlap in the task of entity and relationship extraction..At the same time,by defining the concept of information graph triples,a method for automatically designing information graphs based on information graph triples is proposed.Finally,based on the above work,a text visualization system is implemented.The system can generate the corresponding infographic based on the proportion-related text input by the user,and it also support user-defined design infographic layout,as well as access to design materials through image search engines,which improves the degree of automation and personalization of infographics.