Persistent homology and the upper box dimension
Extremal combinatorics (05D99) Fractals (28A80) Conformal densities and Hausdorff dimension for holomorphic dynamical systems (37F35) Persistent homology and applications, topological data analysis (55N31) Probability measures on topological spaces (60B05) Statistics on metric spaces (62R20) Topological data analysis (62R40)
This paper introduces and investigates a notion of dimension for a metric space defined with respect to the persistent homology of the space. Specifically, the \(PH_i\)-dimension of a metric space \(X\) is defined as follows. We can compute the Čech complex of a subset of the metric space \(\mathsf{x} \subset X\) at distance \(r\), defined as the nerve of the union of balls of radius \(r\) centered at the points of \(\mathsf{x}\). Studying this for a varying parameter \(r\) gives a filtration of simplicial complexes; applying the homology functor to this sequence gives a persistence module and the result can be summarized in a persistence diagram. A persistence diagram, denoted \(PH_i(X)\) for homology dimension \(i\), is a collection of points \((b,d)\) in the upper half plane representing the distances for which a homological feature appeared (``born at \(r = b\)) and disappeared (``dies at \(r=d\)). For this reason, the value \((d-b)\) is often called the ``lifetime of the class represented by the point \((b,d)\). Letting \[E_\alpha^i(\mathsf{x}) = \sum_{(b,d) \in PH_i(\mathsf{x})}(d-b)^\alpha\] the \(PH_i\) dimension is defined to be \[ \dim_{PH}^i(X) = \inf \{ \alpha \mid \exists C \text{ with } E_\alpha^i(\mathsf{x}) <C \; \forall \text{ finite subsets }\mathsf{x} \subseteq X \}. \] The rest of the paper is dedicated to understanding the relationship between \(\dim_{PH}^i\) and the box dimension. In particular, for \(i=0\), \(\dim_{PH}^0(X) = \dim_{box}\). However, the author gives an example of a space where the closely related \(\dim_{\widetilde{PH}}^1\) (which uses Rips instead of Čech complexes) is strictly greater than \(\dim_{box}\). He also shows that for large enough box dimension, \(\dim_{PH}^1 = \dim_{box}\) and conjectures that such a constant exists for higher dimensions \(i>1\).
- A fractal dimension for measures via persistent homology
- Complexity of Delaunay triangulation for points on lower-dimensional polyhedra
- Computational topology. An introduction
- Computing persistent homology
- Coverage in sensor networks via persistent homology
- Delaunay mesh generation.
- Extremal Betti numbers of Vietoris-Rips complexes
- Fractal geometry. Mathematical foundations and applications
- Growth rates of Euclidean minimal spanning trees with power weighted edges
- Hausdorff dimensions for SLE₆.
- scientific article; zbMATH DE number 2103273 (Why is no real title available?)
- scientific article; zbMATH DE number 2115093 (Why is no real title available?)
- Lipschitz functions have \(L_{p}\)-stable persistence
- Measuring shape with topology
- Numerical computations for the Schramm-Loewner evolution
- Persistence stability for geometric complexes
- Persistent homology -- a survey
- Proximity of persistence modules and their diagrams
- Stability of persistence diagrams
- The central limit theorem for weighted minimal spanning trees on random points
- The maximum numbers of faces of a convex polytope
- The minimal spanning tree and the upper box dimension
- The structure and stability of persistence modules
- The Upper Bound Conjecture and Cohen-Macaulay Rings
- Topological persistence and simplification
- Persistence intervals of fractals
- Fractal dimension estimation with persistent homology: a comparative study
- Persistent homology of the sum metric
- A fractal dimension for measures via persistent homology
- 1-Dimensional intrinsic persistence of geodesic spaces
- Alpha magnitude
- Fractal dimension and the persistent homology of random geometric complexes
This page was built for publication: Persistent homology and the upper box dimension
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2227236)