Design And Implementation Of Network Examination System Based On CTT And Reinforcement Learning Theories

With the rapid development and common use of information and network technology these years, network exam comes to be the new trend in the educational testing and evaluation fields. Thus, developing a network examination which much fit for the school courses exam becomes a hotspot for researchers.Against the problems of current network examination systems, the author integrated the theories of classic educational testing and reinforcement learning into the network examination, used the three layers B/S technique architecture together with a series of advance network and database techniques, designed and developed a new network examination system, which can reduce the heavy work of teacher and also can support a good platform of exam, individual practice and learning for students.Firstly, the author studied the item bank design with the classic testing theory, mainly the design of item parameters; studied the strategy of create a scientific test paper considering the reliability and validity; studied how to analysis the examination information from the score, test item and paper. Then based on the reinforcement learning theory, the author designed the individual online practice module, which can let the students make more practice of the knowledge they are wrong in examination, and feel happy with the reinforcement of passing every difficulty.The combination of developing technique can promote the developing efficiency, make the system safe and stable, also can be benefit of the systems extend and maintenance. The system is suit for the school course exam, especially for the courses on language and liberal arts.also it will have important influence on examination pattern and educational reform.