Barcode Posets: Combinatorial Properties and Connections

TL;DR

This paper introduces a new combinatorial framework for analyzing barcodes in topological data analysis, connecting them to permutation-based invariants, poset structures, and polytope geometries, with implications for metric bounds.

Contribution

It defines multipermutation invariants for barcodes, establishes their order relations as crossing numbers, and links these to known posets and polytopes, providing new tools for barcode analysis.

Findings

Posets of barcode invariants are isomorphic to multinomial Newman lattices.

Barcode invariants can bound Wasserstein and bottleneck distances.

The resulting structures form graded face-lattices of new polytopes.

Abstract

A barcode is a finite multiset of intervals on the real line, $B = \{ (b_i, d_i)\}_{i=1}^n$. Barcodes are important objects in topological data analysis, where they serve as summaries of the persistent homology groups of a filtration. The combinatorial properties of barcodes have also been studied, mainly in the context of interval orders and interval graphs. In this paper, we define a new family of maps from the space of barcodes with $n$ bars to the permutation sets of various multisets, known as multipermutations. These multipermutations provide new combinatorial invariants on the space of barcodes. We then define an order relation on these multipermutations, which we show can be interpreted as a crossing number for barcodes, reminiscent of T\'{u}ran's crossing number for graphs. Next, we show that the resulting posets are order-isomorphic to principal ideals of a well known poset known as the multinomial Newman lattice. Consequently, these posets form the graded face-lattices of polytopes, which we refer to as barcode lattices or barcode polytopes. Finally, we show that for a large class of barcodes, these invariants can provide bounds on the Wasserstein and bottleneck distances between a pair of barcodes, linking these discrete invariants to continuous metrics on barcodes.