Computing multidimensional persistence

Authors

  • Gunnar Carlsson Stanford University
  • Gurjeet Singh Stanford University
  • Afra J. Zomorodian Dartmouth College

DOI:

https://doi.org/10.20382/jocg.v1i1a6

Abstract

The theory of multidimensional persistence captures the topology of a multifiltration - a multiparameter family of increasing spaces.  Multifiltrations arise naturally in the topological analysis of scientific data.  In this paper, we give a polynomial time algorithm for computing multidimensional persistence.  We recast this computation as a problem within computational commutative algebra and utilize algorithms from this area to solve it.  While the resulting problem is EXPSPACE-complete and the standard algorithms take doubly-exponential time, we exploit the structure inherent withing multifiltrations to yield practical algorithms.  We implement all algorithms in the paper and provide statistical experiments to demonstrate their feasibility.

Downloads

Download data is not yet available.

Author Biographies

Gunnar Carlsson, Stanford University

Department of Mathematics

Professor

Gurjeet Singh, Stanford University

Department of Mathematics

Research Scientist

Afra J. Zomorodian, Dartmouth College

Department of Computer Science

Assistant Professor

Downloads

Published

2010-11-18

Issue

Section

Articles