Improved EM algorithm for MPT model analysis

Multinomialprocessing tree (MPT) models are in wide use as measurement models for analyzing categorical data in cognitive experiments. The approach involves estimating parameters and conducting hypothesis tests involving parameters that are arrayed in a tree structure designed to represent latent cognitive processes. The standard inference algorithm for these models is based on the well-known expectationmaximization (EM) algorithm. On the basis of the original use of the EMalgorithm for MPT models, this article presents an approach that accelerates the convergence speed of the algorithm by (1) adjusting suitable initial positions for certain parameters to reduce required iterative times and (2) using a series of operations between/among a set of matrices that are specific to the original model structure and information to reduce the time required for a single iteration. As compared with traditional algorithms, the simulation results show that the proposed algorithm has superior efficiency in interpreted languages and also has better algorithm readability and structure flexibility.