A Systematic Comparison of MLE and Bayesian Estimation for MPT Models

Identifier

829

Author

Quan Tang

Date

2013

Document Type

Thesis

Degree Name

Master of Science

Major

Psychology

Concentration

General Psychology

Committee Chair

Xiangen Hu

Committee Member

Arthur Graesser

Committee Member

Seok Wong

Abstract

As a family of statistical models for categorical data, multinomial processingtree (MPT) models have become popular in cognitive psychology over the courseof the past two decades. Classic estimation methods, such as maximumlikelihood estimation (MLE) and model fit test (G2 test), have been applied to MPTmodels widely. Recent development of Bayesian inference suggests a theoreticalalternative for model estimation, though its practical implementation was limiteddue to the difficulties of computation and sampling capacity of the computers. Inthis thesis, I apply Bayesian inference to MPT models, develop the programs thatimplement Bayesian inference for MPT models, and conduct systematiccomparisons between the two approaches in terms of their parameter estimationand model evaluation.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Recommended Citation

Tang, Quan, "A Systematic Comparison of MLE and Bayesian Estimation for MPT Models" (2013). Electronic Theses and Dissertations. 692.
https://digitalcommons.memphis.edu/etd/692