|
Activity Number:
|
129
- Advances in Graph Inference and Network Analysis
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2020 : 1:00 PM to 2:50 PM
|
|
Sponsor:
|
Section on Statistical Learning and Data Science
|
|
Abstract #313273
|
|
|
Title:
|
Online Change Point Detection in Network Sequences
|
|
Author(s):
|
Sharmodeep Bhattacharyya* and Shirshendu Chatterjee
|
|
Companies:
|
Oregon State University and City University of New York
|
|
Keywords:
|
networks;
changepoint detection;
online changepoint;
multiple hypothesis
|
|
Abstract:
|
We consider the problem of change point detection in a sequence of network data. We focus on the online version of the change point detection problem, where the network data arrives sequentially. We develop an algorithm for the detection of both local and global change points in the network sequence. The work introduces a multiple hypothesis testing framework for the detection of change points in a multi-channel setup in the network data context. We also provide a theoretical analysis of the change point detection method for network sequences with each network snapshot generated from an inhomogeneous random graph model. The proposed algorithm detects change points consistently both locally and globally even for sparse networks. We validate and compare our change point estimator using simulated data sets.
|
Authors who are presenting talks have a * after their name.
