Research And Application Of Student Relationship Graph Construction In Campus

With the continuous construction of digital campuses,the campus card has generated massive data,such as students’ basic information,consumption,academics,borrowing and other data.Using data mining technology to analyze students’multidimensional data and mine students’ social behaviors is one of the research hotspots.This paper analyzes the multi-dimensional data of students,mines the hidden social behaviors,and builds a social relationship graph,which is of great significance to the intelligent management of colleges and universities.The main research works are as follows:1.Aiming at the homogeneity in college students’ social behavior,a group hierarchical encounter model based on class association analysis is constructed to eliminate the impact of homogeneity on cross-group social mining.Firstly,we preprocess the consumption data,extract students’ spatio-temporal data,weight the check-in point,and obtain the co-occurrence data based on the sliding window.Secondly,the weight of check-in point is introduced by check-in quantity,and the total data of students is calculated.Secondly,we use the class association analysis with the adaptive threshold to mine the social relations between students within the same group and different groups,so as to construct the student social network.Finally,the experiments verify the effect of cross-group social mining from the professional and grade levels.The experiment shows that our model can improve the cross-group social relationship mining,eliminate the influence of homogeneity on cross-group social mining,and compare the mining results of multiple models that our model is better than other models in terms of recall and precision.2.Based on the rough mining of social relations based on spatio-temporal consumption data,a group hierarchical social driven encounter model based on class association analysis is constructed for fine mining of social relations,to solve the problem that some co-occurrence behaviors may be caused by accidental factors.Firstly,on the basis of consumption data,multi-dimensional data such as academic data,borrowing data and basic information data are added to calculate the similarity of students’ multi-dimensional data;Secondly,combined with Gaussian kernel function,the social driving value of students’ co-occurrence behavior is obtained,and a group hierarchical encounter model based on social driving is constructed to mine the social relations of students from different groups and the same group.Finally,compared with the rough mining model,the experiment shows that the fine mining model is better than the first model in terms of recall and precision.3.Construction the student social relationship graph based on multidimensional data mining.the student social relationship graph system is designed and implemented to display students’ social relations intuitively,facilitate teachers to understand the social dynamics of students with social barriers,guide and help students establish good social relations in time,and make students study and live actively and healthily.