The use of maximum likelihood estimation for evaluate of latent parameters within MPCM model

Dyadkin Yuriy Alekseevich, Bratishchenko Vladimir Vladimorovich

The article is dedicated to Within-Item Partial Credit Model to be a part of Multidimensional Item response theory. The model is considered to be applicable to students’ competences evaluating based on data representing students’ grades on exams. The maximum likelihood estimation is used to get the system of equations that relates latent parameters within the model and empirical data. The numeric values of latent parameters have been achieved by solving the system of equations with the help of Newton's method

Keywords:competences, estimation of latent parameters, Multidimensional Item Response Theory, MIRT, Multidimensional Within-Item Partial Credit Model, maximum likelihood estimation