Abstract Details
Activity Number:
|
254
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 5, 2013 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
Abstract - #308640 |
Title:
|
Sufficient Statistic Selection for Dynamic Networks
|
Author(s):
|
Xizhen Cai*+ and David Hunter
|
Companies:
|
Pennsylvania State University and Pennsylvania State University
|
Keywords:
|
variable selection ;
Cox proportional hazard model ;
penalize partial likelihood
|
Abstract:
|
In recent years, counting processes have been applied to model the dynamics and evolution of large network data. Especially for directed interaction networks, people have been using Cox's model as the underlying intensity process, which allows dependency on a set of sufficient statistics (Perry and Wolfe, 2012, Vu et al 2011). These statistics not only involve measures of time-varying network structure, but also some covariates on nodal characteristics. To include a comprehensive set of sufficient statistics may reduce the bias in predicting future interactions, but also give rise to increase of computational complexity. We first consider networks with single-receiver interactions between nodes, and propose to maximize the penalized partial likelihood to select sufficient statistics and estimate coefficients simultaneously. The resulting estimates of coefficients are sparse and enjoy an oracle property. Interaction networks with multi-receivers are also studied. The approximated partial likelihood is used instead to simplify computations that may bring potential bias. We study the theoretical properties of the maximizer resulting from efficient algorithms.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2013 program
|
2013 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Continuing Education program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.