Spatial-temporal Data Association Based Ontology Alignment Research In High Education Context

During the process of personalized training for university student,there will be large amounts of data through the student activities,which could contain potential regulation of student behavior.Most of these data come from fragmented classes,exams,and lectures and activity context.In order to further extract student behavior patterns,fuse relevant data.On the one hand,we should fully describe the fragmented context.On the other hand,we should solve the data heterogeneity problem during data fusion.Ontology alignment can solve both problems.Context ontology modeling can effectively solve the problem of description of context form,and the ontology alignment process of context ontology reflects the association relationship of the underlying data.Based on the diversity of university context and student behavior and activity data,combining ontology and spatio-temporal data analysis methods,this paper formulates fuzzy reasoning rules to mine the association relationship among the data,so as to achieve context ontology alignment.The main research contents are as follows:(1)Construction of university context ontology : Analyze university context information based on student training programs,and refine relevant concepts and attributes by combining the student data involved in the context to construct the three major context ontology : the class context ontology,the participation context ontology,and the in-depth counseling context ontology.(2)Construction of university context ontology : Part of concepts of university context ontology,such as time,the number of participants,are fuzzy.Therefore,use the fuzzy mathematics method to make concepts and instances of the context ontology fuzzy,and construct fuzzy context ontology.(3)Fuzzy context ontology alignment: Student behavior data from the fragmented college context have time stamp and space stamp which make up the spatio-temporal data,can correspond with activities and behavior in context ontology.These spatio-temporal data correspond to the two concepts of activity and behavior in the context ontology,and their temporal attributes make the two concepts of activity and behavior indistinguishable from the semantic perspective and the data perspective.So this paper proposed a method of context ontology alignment based on time slice algorithm.It will build up the basis of the relationship between activities by calculate the best length range of time slice between activities and behavior,and sort the activities in the time dimension through the formulated fuzzy rules,that is,realize the ontology alignment process through the association relationship between data.(4)Group alignment and case verification: On the basis of achieving individual behavior alignment,a group alignment method based on fuzzy similarity is proposed.According to the actual context and data,construct the fuzzy context ontology to verify the individual ontology alignment and group ontology alignment methods respectively.