Dependence Modeling for Multiway Networks

• Activity Number: 310
• Type: Invited
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: General Methodology
• Author(s): Peter David Hoff*+
• Companies: University of Washington
• Keywords: multivariate ; tucker product ; social network ; separable covariance ; tensor

Abstract

Multivariate networks can be represented as multiway data arrays, potentially having correlations within each set of data indices. For example, a social network may exhibit correlations among the senders and receivers of relationships, or there may be a correlation among the different types of relationships that are measured. In this talk I will develop a class of statistical models for the analysis of such array-valued data. Specifically, I will extend the matrix normal model for matrix-valued data to a class of array normal distributions having separable covariance structure. We relate this model to the higher-order SVD for analysis of array data, and show how the model can be motivated in terms of a latent variable representation. The modeling approach is illustrated in terms of several examples, including social networks and longitudinal, multivariate international relations networks.