Research Of Computer Adaptive Test Based On Irt

Online testing which is an important examination means, freeing people from heavy work of paper examination plays an important role in testing effect. In recent years, online testing has been the focus of academic research.At present, researches on online testing system are considerable home and abroad. Studies mainly focus on, with the assistance of VB ect., exploiting practical applied systems through the analysis of system structure and main function, choice of system developing platform and system network architecture ect. For example:designations of computer testing system based on C/S mode and realization of testing system based on WEB.The testing theories are mostly according to the classical testing theories. These systems are rather practical, but cannot test man individually with regard of their distinctions. While the Item Response Theory does well.Firstly, this paper introduces some concepts and working principles of the Item Response Theory and Computer Adaptive Test.CAT compiling mainly draws to the solutions of the following four questions:ability estimation, topic selection strategy, termination rules and evaluation. Secondly, some common capability estimation methods are introduced in this paper, comparative analysis of their applied situations has been done as well. Then, this paper puts forward the selection methods of weighted in terms of topic selection strategies. Errors of new topic selection strategy are calculated by computer simulation calculation, making comparative analysis with the results in literature. Experimental results show that the application, blasting luminosity, efficiency and deviation of new scheme have great advantages, Improving the objectivity and accuracy of the system testing results, reaching the goal of testing individually according to distinctions well.