Advisor
Schaffer, Jay Ryan, 1969-
Advisor
Lalonde, Trent L.
Committee Member
Franklin, Scott B.
Committee Member
Shafie, Khalil
Department
Applied Statistics and Research Methods
Institution
University of Northern Colorado
Type of Resources
Text
Place of Publication
Greeley (Colo.)
Publisher
University of Northern Colorado
Date Created
5-1-2014
Genre
Thesis
Extent
175 pages
Digital Origin
Born digital
Description
Binary data that are correlated across space and time often occur in health and ecological studies. The centered spatial-temporal autologistic regression model (Wang & Zheng, 2013) accounts for the spatial and temporal dependence that can occur in binary data. Statistical inference for the autologistic model has been based upon pseudolikelihood, Monte Carlo maximum likelihood, Monte Carlo expectation maximization or Bayesian hierarchical models. However, these methods require the full conditional distribution to be defined and with the complexity of spatial and temporal dependence and interactions between observations can cause convergence problems and increase computation time. An alternative approach to likelihood based methods for spatial-temporal data is to use generalized method of moments, a method not currently used for spatial-temporal binary data. Two different generalized method of moments approaches based on a set of moment conditions is constructed with respect to spatial neighborhoods to account for the spatial and temporal dependence of the data. Comparisons of estimation methods using a small simulation and two data sets show that the generalized method of moments approaches perform well in specific data situations.
Notes
Dean's Citation for Excellence
Degree type
PhD
Degree Name
Doctoral
Language
English
Local Identifiers
Kaufeld_unco_0161D_10314
Rights Statement
Copyright is held by author.
