Graph Estimation from Multi-Attribute Data

• Activity Number: 190
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Mladen Kolar*+ and Han Liu and Eric P. Xing
• Companies: Carnegie Mellon University and Princeton University and Carnegie Mellon University
• Keywords: Graphical model selection ; Multi-attribute data ; Network analysis ; Partial canonical correlation

Abstract

Many real world network problems often concern multivariate nodal attributes such as image, textual, and multi-view feature vectors on nodes, rather than simple univariate nodal attributes. The existing graph estimation methods built on Gaussian graphical models and covariance selection algorithms can not handle such data, neither can the theories developed around such methods be directly applied. In this paper, we propose a new principled framework for estimating multi-attribute networks. Instead of estimating the partial correlation as in current literature, our method estimates the partial canonical correlations that naturally accommodate complex nodal features. Computationally, we provide an efficient algorithm which utilizes the multi-attribute structure. Theoretically, we provide sufficient conditions which guarantee consistent graph recovery. Empirically, we apply our method on a genomic dataset to illustrate its usefulness.