Griffin and Li: Unsupervised Learning of Dyadic Processes: Models, Methods, and Simulation

William A. Griffin and Xun Li

The dynamics that arise from dyadic processes, such as those observed in married couples, generate a cascade of effects- some good and some bad- on each partner, other family members, and other social contacts. Although the effects are well documented, the processes associated with the varied outcomes are not well understood. We currently have two methods of simulating dyadic interaction in married couples- an algorithm-based ABM and a particle filter. Although both work reasonably well, neither fully captures the dynamics of an evolving social process as well as we would like. Recently, we developed a third method of generating couple dynamics model using a Hierarchical Dirichlet Process Hidden semi-Markov Model (HDP-HSMM) (Johnson & Willsky, 2012). We review the how this technique generates plausible dyadic sequences that are sensitive to relationship quality and provides a natural mechanism for simulating micro-social processes.

Download PDF.

Leave a Reply