Learning Analytics Research Based On The OIM Relationship Chain

With the innovation and development of information and communication technology,human society begins to enter the era of big data.As the main application of big data in the field of education,learning analytics has become a new coming wave leading the development and reform of education information.It has important strategic value.However,as learning analytics gradually moves from theoretical research to practical exploration,its application in the school and even the entire education field is still very limited.There are many practical difficulties and challenges in data,technology,system,management,privacy,and so on.Therefore,the Objective/Indicator/Method(OIM)relationship chain is introduced in this thesis.The main purpose of this thesis is to explore the general rules of learning analysis practice by multi-case analysis,establish some OIM relationship chains for different problem scenarios of online learning platform,and build an OIM-based learning analysis system architecture.This thesis provides technical and methodological guidance for the practice of learning analysis in the future,and enriches the design theory of learning analytics system architecture.First,this thesis systematically reviews the classic foreign learning analysis model,combs the application of pedagogical design ideas in the learning analysis practice,and on this basis,summarizes the common analysis objectives,indicators,methods in learning analysis.Second,in this thesis,144 learning analysis application cases are selected from the International Conference on Learning Analysis and Knowledge(LAK)in the last five years,and analyzed by content analysis and case meta-research in three dimensions of objectives,indicators and methods.Based on the results of cluster analysis of 17 quantitative indicators,this thesis establishes 10 OIM relationship chains for different problem scenarios of online learning platform,such as modeling learner knowledge,predicting learning performance.Then,referencing to the typical foreign learning analysis system architecture,this thesis constructs an open learning analysis system architecture based on OIM relationship chains.This architecture includes 9 major core components of data management,data mapping,OIM relationship chain,privacy rules,learner portraits,context modeling,analytics modules,visual presentation,and analytics engine.Finally,the OIM relationship chains and the open learning analytics system architecture are evaluated through interviews of expert and technicians.The results of the evaluation show that the OIM relationship chains and the system architecture constructed in this thesis are scientific,available,and effective.At the same time,this thesis uses the "Identification for high-risk student" chain as an example to design a "learning risk warning" case based on the Shanghai X University online learning platform,which provides a demonstration of the application of the OIM relationship chain and further validates the availability of the relationship chain.