Analysis And Prediction Of College Students’ Social Networks From The Perspective Of Psychology

Social intercourse is an essential part of college students’ daily life,and lots of important information will be hidden in the students’ social networks.Exploring the characteristics of college students’ social networks and predicting the friendship can discover the social rules of students.It also can detect students with abnormal and extreme behaviors in advance,and provide a reference for college managers.Therefore,this paper mainly analyzes the characteristics of college students’ social networks from the perspective of psychology,and proposes a social network centrality prediction method and a friendship prediction method.At present,scholars have done some researches on the analysis and prediction of college students’ social networks,but little to no work has combined psychology and big data methods.Therefore,it is still a challenge to analyze and explore the social regularity of college students from a psychological perspective.First,this paper builds social networks in different social scenarios based on psychology.Then,based on statistics and network science we use network science indicators to explore the rules in networks and the factors that affect network centrality.At the same time,this paper analyzes the psychological characteristics and facial perception information of college students in social networks,then proposes a framework named SEASON(Social Situation Prediction Based on Facial Perception and Psychological Information)to predict the position of college students in social networks under different social situations.Finally,this paper proposes a framework named DEFINE(Node Enhancement based Friendship Detection)based on the college student friendship network scenario.This framework expands the network representation learning algorithm enhanced by node attributes,which combine students’ facial perception characteristics,psychological characteristics and network structure characteristics to predict the friendship between college students.The centrality prediction method of college students’ social network and the prediction method of friendship relationship proposed in this paper have carried out a large number of experimental analysis on real college student social networks.Experimental results show that SEASON can more effectively predict the centrality of college students in different social networks,and DEFINE can more effectively predict the friendship of college students.