Saved in:
Bibliographic Details
Main Authors: Krebs, Johannes, Rademacher, Daniel
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2311.11259
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916426461216768
author Krebs, Johannes
Rademacher, Daniel
author_facet Krebs, Johannes
Rademacher, Daniel
contents In this paper, we study the stability of commonly used filtration functions in topological data analysis under small perturbations of the underlying nonrandom point cloud. Relying on these stability results, we then develop a test procedure to detect and determine structural breaks in a sequence of topological data objects obtained from weakly dependent data. The proposed method applies for instance to statistics of persistence diagrams of $\mathbb{R}^d$-valued Bernoulli shift systems under the Čech or Vietoris-Rips filtration.
format Preprint
id arxiv_https___arxiv_org_abs_2311_11259
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On the stability of filtration functions for dependent data with applications to break detection
Krebs, Johannes
Rademacher, Daniel
Statistics Theory
In this paper, we study the stability of commonly used filtration functions in topological data analysis under small perturbations of the underlying nonrandom point cloud. Relying on these stability results, we then develop a test procedure to detect and determine structural breaks in a sequence of topological data objects obtained from weakly dependent data. The proposed method applies for instance to statistics of persistence diagrams of $\mathbb{R}^d$-valued Bernoulli shift systems under the Čech or Vietoris-Rips filtration.
title On the stability of filtration functions for dependent data with applications to break detection
topic Statistics Theory
url https://arxiv.org/abs/2311.11259