Change-Point Detection for Multivariate and Non-Euclidean Data with Local Dependency

• Activity Number: 140 - Change-Points in Multivariate and High-Dimensional Data
• Type: Invited
• Date/Time: Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Hao Chen*
• Companies: University of California, Davis
• Keywords: non-parametric; graph-based test; tail probability; high-dimensional data; network data

Abstract

In a sequence of multivariate observations or non-Euclidean data objects, such as networks, local dependence is common and could result in false change-point discoveries. In this work, we study a new way of permutation, circular block permutation with a random starting point, on a graph-based change-point detection framework, leading to a general framework for change-point detection on data with local dependence. We provide analytic formulas to approximate the test statistic and the circular block permutation p-value, making the new framework an easy off-the- shelf tool for data analysis.