Semantic Portrait Of Social Relationship Based On User Portrait

With the rapid development of social media in the current Internet environment,a huge amount of users' electronic footprint information can be obtained from these resources,which contains rich data of people's social speech,social relations,social behavior,etc.The data provides us with necessary support so that it is conducive to building user portraits and analyzing social relationships.Social relationship is an important part of human society,and is essential to the study of the social network and social computing.Traditional social relationship analysis mostly tends to discussing the direction,intensity and type of the relationship,which lacks rich semantic connotation.Therefore,the relationship portraits constructed by aforesaid method has limitations and it is difficult to provide enough support to advanced social network study.This paper proposes to construct user portraits with rich semantic connotation through the massive data of the social media taking word vector technology,keyword extraction technology and psychological theory.Semantic relationship portraits will be constructed from three perspectives based on user portraits by means of non-intuitive pattern,intuitive pattern and Big Five personality theory.Based on the analysis of the experimental results of the above three perspectives,we find the advantages and disadvantages of three methods respectively.For example,the non-intuitive way is obviously flawed in performance connotation,but it is convenient to calculate;In the intuitive way,the relationship connotation can be judged by human subjectivity,but the calculation is difficult.Finally,a visualization system on user relationship portraits is developed.The system can display the macroscopic social network,the user portraits and the relationship portraits.It can also conduct the screening of the relationship network under certain conditions and the calculation of the relationship path then demonstrate on the Sina Microblog dataset.