Adaptive Testing Model Approach Based On Birnbaum Model And Markov Model

Computerized adaptive testing is widely used in the testing and assessment of the level of learners' competency. Popular computerized adaptive testing systems now use the mathematical models of Item Response Theory based on the relationship between the ability of examinees and item parameters. However, Item Response Theory does not take into account the impact between previous answers and the next item selection. The item parameters principally rely on probability methods of Classical Test Theory. This article proposes the combination of Item Response Theory (Birnbaum model) and Markov chain to calculate the dependency of answer set during the quiz process. Concurrently, using Hooke - Jeeves direct search method within the limited range of parameters to assess the set of item parameters.