Research & Implementation Of Computerized Adaptive Testing System

Traditional computer-aided test focus on the time and numbers of test but ignores the differences between testers, which may leads a waste of time and defeats the tester's confidence. Computerized adaptive testing (CAT), based on item response theory (IRT), can adapt the tester's characteristics initiatively, extract test questions dynamically, shorten the testing time and improve the accuracy of the testing results.Being study on the related theory to CAT and technology of the system implementation, the paper designs a CAT for history of junior school and implements it. There are four key problems to solve, item parameter, item-selected algorithm, ability estimating algorithm and termination strategy. Base on problems, this paper works on: for the item parameter, uses sample statistic and joint maximum likelihood estimation algorithm, to get the initial value of the item parameters and correct it. For the item-selected algorithm, uses optimizes maximum information item selection algorithm to reduce the amount of calculation of the algorithm. For ability estimation algorithm, uses condition maximum likelihood estimation algorithm and joint maximum likelihood estimation algorithm to improve the accuracy of the testing results under the conditions that item parameters are uncertain. For termination strategy, reduces the testing time by using fixed testing time algorithm and no fixed testing time algorithm.The main research content of this paper is: first part explains the key theory about item response theory (IRT), the key algorithms of computerized adaptive testing (item-selected algorithm, ability estimating algorithm and termination strategy and the relevant J2EE technology for the implementation of the system). Above the key algorithms, this paper produces the concrete realization method. At the second part, this paper elaborates on the design of this system, design and implementation of each function module. Analysis and test the system, in order to verify the performance of the proposed method. Finally, summarizes the advantages and disadvantages of the system and gives an outlook about the development prospects of computerized adaptive testing.