The Research And Design Of Test Paper Composition Algorithm Adaptive Testing System

With the development and popularization of computer and network technology, Computerized examination has been applied more and more extensive.This form of examination can reduce the human subjective factors in the examination,it also can organize the examination more quickly,more conveniently,and more accurately,most significantly,it can reduce the workload of the teachers.At present,computer-based test has been widely applied to various education examinations,title examination and certification examinations.But most of the examination systems only change the traditional way of manual questions into computer questions,the basic idea has not broken through the traditional classical test theory,and it can not reflect the true level of the examinee well.In order to assess the true level of examinee more accurately,the Computer Adaptive Test which based on the Item Response Theory treats the examinee as a object,according to different examinee who reflect the ability and level in the examination,adjust the difficulty of the examination dynamically.The paper firstly research the background and status of the CAT system,discusses the theoretical basis for CAT exam-CAT,elaborate several methods of parameter estimation of examinee's ability.And then it determines the method of the ability value estimation and the initial subject selection strategy,and put forth a design idea and solution for CAT system.Test paper composition algorithm is the focus of this paper, through the analysis and comparison of the various strategies about topics selection,the paper bring forward a optimized method about topics selection,which is based on feasibility and practicality.It not only can estimate the initial ability of the examinee, but also can compete the self-adaptive chosen.Finally,in order to verify the feasibility and practicality of the system,the paper test the system,record and analyze the experimental results for future improvement and perfection.