Research And Implementation Of Adaptive Test And Test Item Recommendation System

Today,with the rapid development of Internet education,online learning has become an important learning method.Learning can be carried out anytime and anywhere,and learners can obtain different online learning resources in various ways.However,the explosive growth of learning resources generated daily on the Internet has caused learners to face problems such as "learning trek".In order to solve these problems,adaptive test systems and recommendation systems have gradually become research hotspots in recent years.The adaptive testing system selects the knowledge points that need to be learned next according to the user's current knowledge state,which is in line with people's cognitive process from simple to complex,but the learning path and learning time of students are relatively fixed,ignoring the personalization of people.The recommendation system can predict the preferences of other resources that have not had interactive behaviors based on the user's behavior,thereby recommending personalized learning resources suitable for students,but it is difficult for the recommended learning resources to comply with the relationship between knowledge content.This paper studies the theory of knowledge space and the principle of recommendation system,combining the advantages of the two,develops an adaptive test and test recommendation system,and solves the problem that in the process of learning,it not only follows the cognitive laws of learning,but also meets the personality of learners with different knowledge bases.To meet the requirements of learning,avoid "problem sea tactics".The main work carried out in this paper is as follows:1.Research the related theories of knowledge space.Two strategies for solving the boundary of knowledge state are analyzed,namely,the selection strategy of knowledge state boundary based on test questions and the selection strategy of knowledge state boundary based on knowledge points,and the advantages and disadvantages of the two are compared.2.Design a topic selection scheme based on latent factor model.Use the score data of users to answer test questions for model training,mine potential user characteristics and test questions characteristics,and predict the user's score for unanswered test questions.Three methods for selecting Top-N test questions in different ordering directions are proposed to select test questions of different difficulty for users of different ability levels.3.Propose an algorithm for filtering recommended test questions based on the knowledge space theory.According to the answer test results,based on the knowledge point labels of the test questions,the candidate's knowledge state is approximated.And based on the boundary of the knowledge state,select the next set of knowledge points to be tested,and accurately filter the test questions recommendation.The test to determine the knowledge domain can effectively shorten the test step.4.Develop an adaptive test and test question recommendation system.The main functions of the system include adaptive question selection module,random question selection module,wrong question correction module,test question collection module,etc.Experiments show that this system can not only be tested in sequence according to the cognitive process,but also can be recommended to learners with equivalent test questions.For learners with a certain foundation,it can effectively shorten the test step.