Research And Implementation Of An Adaptive Online Learning Testing System

Along with the rapid development of the Internet,the education field has undergone tremendous changes.The emergence of online education expands the boundaries of traditional education and benefits every student,while also promoting the sharing of educational resources and more comprehensive educational assessment theories and techniques.Various online learning systems such as MOOC provide online learning services for people.However,various existing online learning systems have more or less defects: there is no individualized learning program for individuals,and the classification of knowledge is relatively coarse,which leads to the learner not being able to complete the study well and the physical examination of the study is poor.In order to enable learners to have a reference learning path and better assess the learner's mastery in the learning process,this thesis studies and implements an adaptive online learning testing system.The main work of the thesis is divided into three aspects:Firstly,an efficient learning strategy based on Q learning algorithm is proposed.Firstly,the hierarchical knowledge is used to establish the subject knowledge as a gridlike knowledge map.Then,the Q learning algorithm in reinforcement learning is used to construct an efficient learning path.Finally,the learner learns according to the given learning path;Secondly,proposed an improved topic selection strategy.By introducing the threshold limit of exposure rate,the improved topic selection strategy makes the test questions in the question bank more uniform,and the questions of different difficulty levels may be selected,and the security of the test bank is guaranteed;Thirdly,design and implement an adaptive online learning assessment system.Based on the consideration of practicability and feasibility,the online learning system architecture was designed and the key modules of the system were realized.The hierarchical learning model is used to complete the knowledge map of the Data Structure course,and then the efficient learning path of the Data Structure course is constructed based on the Q learning algorithm,and describe in detail the learning path of the queue knowledge unit,and finally propose an improved topic selection strategy.In the simulation experiment of the thesis,we compare the improved topic selection strategy with the traditional topic selection strategy such as the maximum Fisher information method and the a layering method.The results show that the improved topic selection strategy is superior to the traditional topic selection strategy in the comparison of evaluation indicators.Through the combination of the efficient learning path and the improved topic selection strategy,an adaptive online learning evaluation system is designed and implemented.