|
Activity Number:
|
92
- Time Series and Finance
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 9, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
|
Abstract #317826
|
|
|
Title:
|
Community Network Auto-Regression for High-Dimensional Time Series
|
|
Author(s):
|
Elynn Y. Chen* and Jianqing Fan and Xuening Zhu
|
|
Companies:
|
University of California, Berkeley and Princeton University and Fudan University
|
|
Keywords:
|
Network autoregression;
Community structure;
Common latent factors;
High-dimensional time series;
VAR model
|
|
Abstract:
|
Modeling responses on the nodes of a large-scale network is an important task that arises commonly in practice. This paper proposes a community network vector autoregressive (CNAR) model, which utilizes the network structure to characterize the dependence and intra-community homogeneity of the high-dimensional time series. The CNAR model is fundamentally different from the network vector autoregressive NAR (Zhu et al. 2017) model in that the CNAR is a population-level generative model while the NAR only utilizes a realized sample adjacency matrix. The CNAR model greatly increases the flexibility and generality of the NAR model by allowing heterogeneous network effects across different network communities. In addition, the non-community-related latent factors are included to account for unknown cross-sectional dependence. The number of network communities can diverge as the network expands, which leads to estimating a diverging number of model parameters. We obtain a set of stationary conditions and develop an efficient two-step weighted least-squares estimator. The consistency and asymptotic normality properties of the estimators are established.
|
Authors who are presenting talks have a * after their name.
