Research On Selection Method Of Cognitive Diagnosis Adaptive Test Based On Knowledge Space Theory

With the acceleration of information construction in education-related industries,various online learning platforms such as online learning space and MOOCs emerge in an endless stream.Individualized learning assistant services have become the focus of attention of various online learning platforms.Individualized teaching application services such as intelligent push of learning resources,learning path planning and learning situation analysis are indispensable.The realization of these functions requires the acquisition of relevant data such as students'learning habits,learning abilities and students' cognitive status.The data related to learning habits and learning abilities need to rely on daily learning behavior records and analysis,which can help students solve the problem of "how to learn".To solve the problem of "what to learn" for students,we need to depend on the students'knowledge state,which depends on the test means with co gnitive diagnostic function.Common computer testing methods are generally based on classical testing theory.It is still a fixed topic,fixed time and fixed order to use computers instead of writin g tools to complete traditional paper and pencil testing.Very little attention is paid to the tester's personal status,and it is difficult to meet the needs of personalized learning.Therefore,computer adaptive testing technology is gradually developing.Nowadays,adaptive testing based on item response theory is very mature.However,due to its purpose of inspecting the ability of the tested person and paying insufficient attention to the knowledge state of the tested person,the computer adaptive cognitive diagnostic testing method aiming at acquiring the knowledge state of the students has come into the eyes of many experts and scholars.The measurement method with diagnostic ability can not only inspect students'ability level,but also inspect students' cognitive structure status.These test results can provide important data support for online learning platform resources push or personalized learning guidance and other educational services.Cognitive diagnosis adaptive testing technology is facing the problems of single question bank construction,complex topic selection methods,imperfect test termination methods,and so on,so it has not been widely popularized.As the core of the adaptive test system,how to select the test questions efficiently and reasonably becomes the research focus.Common topics selection methods such as KL-based information quantity selection method and knowledge space theory-based topic selection method have complex calculation process.If the question bank and the cognitive attribute with test knowledge domain are large and complex,the calculation amount will be extremely hu ge and difficult to apply.In the third chapter,based on the theory of knowledge space and combined with the technology of engineering state search,a hierarchical topic selection strategy for item bank is proposed.The core idea of this strategy is to lay the question bank according to the number of knowledge attributes and the structural relationship between knowledge attributes,and then according to the actual project response of the subjects.The state dynamically adjusts the test question bank.Item selection is based on the temporary knowledge state of the subject,and the highest degree of support for the project attributes is the best choice.Using the least test questions to cover the largest range of attributes as far as possible can effectively improve the test efficiency under the condition that the test results are accurate.Chapter 4 of this paper develops the test system to verify the feasibility of the adaptive test system applied to the online learning platform.