Seminar Details

Seminar Details


Feb 6

3:30 pm

Latent Factor Models for Relational Data

Peter Hoff


University of Washington - Department of Statistics & Biostatistics & the Center for Statistics and the Social Sciences

Matrix representation techniques have a long history in the analysis of multivariate data, including relational data in which observations are on pairs of individuals or units. In particular, the singular value decomposition of a matrix allows one to represent the relationship between two units as the inner product of a pair of latent characteristic vectors. In this talk I discuss a model-based version of the singular value decomposition which allows for the analysis of a variety of data types, including binary relational data, or “social networks.”

One outstanding issue in the use of such models has been the determination of the dimension of this latent space of characteristics. Time permitting, I will show how Bayesian methods can be used to select an appropriate dimension, and how Bayesian model averaging over the dimension can improve upon the predictive power of these models.

Background papers: