Workshop Series: Andre Bonfrer
Abstract: The sociological theory of homophily states that people who are similar to one another are more likely to interact with one another. This can be translated into a stochastic statement about a probability-based link between similarity and interactions. Frequently marketers have access to data on interactions among customers from which, with homophily as a guiding principle, inferences can be made about underlying similarities among. In this paper we extend the latent space modeling framework of Handcock and Raftery (2007) to infer similarities among customers, from observations of interactions among these customers. In particular, we present a novel semi-parametric Bayesian approach, using Dirichlet processes, to moderate the scalability problems that marketing researchers encounter when working with networked data. We find that this framework is a powerful way to draw insights into latent similarities of customers, and we discuss a number of ways this can be useful to marketers. We also focus this framework on the task of predicting who will contact whom, and demonstrate significant predictive performance advantages over several commonly cited link mining algorithms used for this task.
Join our staff, students and alumni attending workshops presented by visiting academics on their area of research expertise.