Research On Multi-dimensional People Relationship Modeling And Visualization Method

Paying attention to people relationships is one of the main characteristics of traditional Chinese culture.Some scholars even use people relationships to define people themselves.Due to the complexity of the people relationship,mining and displaying the hidden rules among people by data mining,graphic visualization technologies has become one of the research hotspots in the computer domain.However,in society,people's identities are diverse,and the relationship between people is also multi-dimensional.The existing modeling methods of people relationships do not consider the multidimensionality of social relationships.These methods make the current social relationship visualization only available for the relationships among microcommunity,difficult to fully reveal social relationships,which leads to the one-sided issue of people relationships analysis.This paper mainly studies the modeling and visualization for multi-dimensional people relationships,which has two aspects: modern people relationships and historical people relationships.The specific research contents are as follows:(1)A photo-based visual analysis method for social relationships is proposed.Firstly,face recognition technology is applied on social photos to build a strong social relationship based on the co-occurrence of faces,that is,there must be intersections between people in the real world.Secondly,the microblog data of the identified people is obtained through a microblog crawler,and a virtual social relationship is added on the strong social relationship through an analysis of the microblog data.To model more comprehensive social relationship,this paper adds the implicit social relationship,and uses the LDA theme model to measure of the similarity between users.Finally,various visualization methods are presented for different types of social relationships.(2)A multi-dimensional visual analysis method for the relationship of historical people is proposed.Based on the data of Chinese Historical Biographical Database(CBDB),a new visual analysis method of historical people combining social relations,kinship relations and geographic information was proposed,which implemented for five dynastyies,including Tang,Song,Yuan,Ming and Qing.For the visualization part,we employed a combination of pie charts,sanghi charts,and scattered points instead of network charts to display the relationship of historical people,and used tree charts to present the genealogies and the master-and-apprentice relationship.The purpose of this work is to use computer technology to provide the users an intuitive,more acceptable and understandable mode for complex and diverse people relationships.Our method is able to model multi-dimensional people relationships,and combines a variety of visualization methods.Through a large number of user surveys,the visual analysis method combining real social and virtual social proposed in this article provides a more adequate and reliable friend recommendation and social network analysis.A variety of people relationships are combined with geographic information to visualize the multi-dimensional relationship of historical people,which provides a new perspective for the research and exploration of humanistic domain.