B1204
Title: Learning joint and individual structure in network data with covariates
Authors: Jesus Arroyo - Texas A&M University (United States) [presenting]
Carson James - Texas A and M University (United States)
Dongbang Yuan - Texas A and M University (United States)
Irina Gaynanova - Texas A and M University (United States)
Abstract: Datasets consisting of a network and covariates associated with its vertices have become ubiquitous. One problem pertaining to this type of data is to identify information unique to the network, information unique to the vertex covariates and information that is shared between the network and the vertex covariates. Existing techniques for network data focus on capturing structure that is shared between a network and the vertex covariates but are not able to differentiate structure that is unique to each. A solution is formulated via a low-rank model and a two-step estimation procedure, composed of an efficient spectral method to obtain an initial estimate for the joint structure, followed by an optimization method that minimizes a nonconvex loss function associated with the model. The consistency of the initial estimate is studied, and the performance on simulated and real data is evaluated.