Parametrized homology via zigzag persistence

Abstract: This paper develops the idea of homology for 1-parameter families of topological spaces. We express parametrized homology as a collection of real intervals with each corresponding to a homological feature supported over that interval or, equivalently, as a persistence diagram. By defining persistence in terms of finite rectangle measures, we classify barcode intervals into four classes. Each of these conveys how the homological features perish at both ends of the interval over which they are defined.

Review text

Let \(X\) be a topological space, and let \(f: X\to \mathbb{R}\) be a real-valued function. This paper introduces \textit{parametrized homology} to study the homology groups of the fibers \(f^{-1}(r)\) of \(f\), and also the relationship between different fibers of \(f\), as the parameter \(r\in \mathbb{R}\) varies. For \(r\neq r'\), the fibers \(f^{-1}(r)\) and \(f^{-1}(r')\) are disjoint, and there are no obvious maps from one fiber to another. Therefore, it is not easy to associate a lifetime of parameter values to a homological feature present in any fiber. Nevertheless, parametrized homology allows one to do exactly that -- associate a birth value and a death value to any homological feature in the fibers, in analogy with persistent homology. The authors define parametrized homology using rectangular measures, and prove many of its properties using an attractive graphical notation for quiver representations or zigzag modules. They show how in finite situations (the discrete Morse-like case), parametrized homology recovers zigzag persistent homology. Indeed, parametrized homology should be thought of as a way to build a variant on zigzag persistence that allows for a continuous parameter; it is not at all obvious how to do this. The authors also prove a stability theorem: If \(f: X\to \mathbb{R}\) and \(g: X\to \mathbb{R}\) are sufficiently nice spaces and maps, then the distance between the parametrized homologies for \(f\) and \(g\) is bounded from above by the sup-norm distance between the functions \(f\) and \(g\). Parametrized homology (a) recovers the homological information stored in \(f: X\to \mathbb{R}\), (b) is symmetric with respect to replacing \(f\) with \(-f\), and (c) does not require strong finiteness assumptions. This distinguishes parametrized homology from sublevelset persistent homology, which does not fully satisfy (a) or (b). Zigzag homology requires one to discretize the real line into subintervals, stubbornly failing (c). Extended persistent homology is closely related to parametrized homology, and indeed the authors show a one-to-one correspondence between them. Nevertheless, the symmetry (b) in extended persistent homology is complicated, requiring one to consider multiple homological dimensions at once.

Recommendations

Cites work

Cited in

(17)

This page was built for publication: Parametrized homology via zigzag persistence

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2000778)