Factor Models for High-Dimensional Dynamic Networks: With Application to International Trade Flow Time Series 1981--2015

• Activity Number: 132 - Statistical Analysis for Networks
• Type: Contributed
• Date/Time: Monday, July 30, 2018 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Elynn CHEN* and Rong Chen
• Companies: Rutgers University and Rutgers University
• Keywords: Dynamic Network/Relational Data; Matrix-Variate Time Series; Factor Model; Eigen-Analysis; Convergence in L2-norm; Dimension reduction

Abstract

Dynamic network analysis has found an increasing interest in the literature because of the importance of different kinds of dynamic social and economic networks. Different from current methods that characterize the dynamic networks on the node and edge level, our approach treats the evolving sequence of networks as a time series of squared relational matrices. We adopt a matrix factor model where the observed surface dynamic network is assumed to be driven by a latent dynamic network with lower dimensions. The proposed method is able to achieve dimension reduction and to unveil the latent dynamic structure. Different from other dynamic network analytical methods that build on latent variables, our approach imposes neither any distributional assumptions on the underlying network nor any parametric forms of its covariance function. The latent network is learned directly from the data with little subjective input. We applied the proposed method to the monthly international trade flow data from 1981 to 2015. The results reveal an interesting evolution of a trading network among latent groups and relations between the latent trading groups and the countries.