|
Testing is a widely taken method in teaching evaluation. Along with the decades ofrapid development of computer and internet techniques, and for the teaching evaluationthere also comes up new realization mode, which is to apply computer as testing tool intoteaching evaluation. There are two modes when use computer to run tests, they areComputer-Based Tests, CBT and Computerized Adaptive Testing, CAT. The former basedon the Classic Testing Theory, use computer to manage and score the test, usually beencalled Computerized Fixed Tests. What’s more, more and more tests are internet oriented,and therefore called Internet Based Test (IBT). However, same to PPT, when implementCBT, all the tested are shown the same number and same sequence of questions, but seldomor never pay attention to the individual ability of the tested.Different with CBT, the idea of CAT is to simulate the intelligent evaluation of human,and by the choice of computer, every student is given suitable test item according todifferent ability level, and to do more accurate evaluation to the ability level of the testedthrough statistics of the response of the tested to the item. This more flexible test solves theproblems of the traditional test, prevents the students from answering too many complicatedor simple items. Theoretical analysis and practical result shows that when the selected itemis suitable for the tested ability level, the tested is most motivated, and the reliability andeffect of the test is the best. Thus, CAT could truly achieve to run test according to ability.Compared to the traditional PPT or CBT, the advantages of CAT includes1）more accurateevaluation of the intellectual level of the tested;2）decrease the test length and timeremarkably;3）improve the learning motive of the tested;4）the most important thing is, useCAT to do the evaluation could be more suitable to the actual ability of every individualtested. In abroad, the CAT study has been considerable progress, and has been widelyused in many fields.The domestic researches about CAT starts relatively late. Thus, theresearch for CAT has important theoretical value and practical meaning.As the research objective is the problem of CAT as well as the technology, and theresearch of this paper mainly includes the following aspects:（1）Through the introduction of the examination’s development process,basic theory,psychology basis and basic components of CAT, and reviews, analyses and summarizes theresearch status quo of present domestic and aboard. Achieve the purpose of accurateevaluation of the candidates.（2）It does a detailed analysis of the basic hypothesis in the item response theory. Aimat the disadvantages of the model in item response time, in the first, this paper proposed item response time hypothesis, and adds the item response time to the item response model,proposes the Logistic-T model and deduces and certify this model theoretically.（3）It explains the necessity of the item parameter amendment in ComputerizedAdaptive testing, and the implement problems of commonly used amendment method, andproposes the item parameter amendment method according to item usage information.（4）It introduces and does comparative analysis of the main item selection method atpresent, points out the problems in the application. Then, it introduces the design principlesof item selection method, and on this basis, it proposes new item selection method. In themean time, in order to improve the efficiency of item selection, it adopts simulatedannealing algorithm to solve the target function, form the TIS-SA item selection method andachieve relatively good effect.（5）At last, it realizes the adaptive testing prototype system CAT-GD,and realizes thekey technologies which proposed in this study,such as Logistic-T item response model, itemcalibration method b-CBUI, TIS-SA item selection engine, etc, and does the testing andcomparative analysis.
|