Bilgisayarda doğrusal olmayan denklem modellemesi bilgisayar ortamlı sınama

Computer adaptive test (CAT) designed to estimate the ability of a student, where a collection of questions that has different level of difficulty are selected according to the level of student's ability. CAT used item response theory (IRT) as a model to select the proper questions and to estimate the ability of the student. Aiming at the problems in estimation of the ability of the student in IRT, we suggest two methods which are, random walk theory and hidden Markov model (HMM). We conduct a simulation experiments for one-dimensional random walk models and HMM model. The result shows that the one-dimensional random walk models and HMM model are less complex and better than IRT model.