←Back
686 – Bayesian Modeling in the Social Sciences and Education
Understanding Variability Between Groups of Sequences Using a Bayesian Object-Oriented Data Model
Sponsor: Section on Bayesian Statistical Science
Keywords: Bayesian Methods, Life course data, object-oriented data, Optimal Matching, Sequence analysis, Social Sciences
Maria Tackett
University of Virginia
Dan Spitzner
University of Virginia
Optimal Matching is an algorithm used to analyze object-oriented data, specifically sequences. In this article, we propose a framework inspired by classical Analysis of Variance to estimate the variability between groups of sequences using Optimal Matching. To estimate between group vari- ability, we explore approaches based on established methods- bootstrapping and Hidden Markov Models. We also propose a Bayesian object-oriented model. These three methods are examined and demonstrated on a well-known dataset of historical English dance sequences, with a mind towards application to life course data.