A graph-based model to discover preference structure from choice data

AbstractIn this paper we demonstrate how to use graph matching to uncover heterogeneity in the structure of preferences across a population of decision-makers. We propose a novel non-parametric approach to formally capture the concept of preference structure using preference graphs, thereafter clustering decision-makers based on graph embedding methods. We explore the approach with simulated choice and empirical data from the most common classes of economic and psychological models. The approach uncovers heterogeneity in preference structure across a variety of dimensions, without requiring any prior knowledge of those structures.

Return to previous page