Activity preserving graph simplification

Francesco Bonchi, Gianmarco Morales, Aristides Gionis, Antti Ukkonen

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

We study the problem of simplifying a given directed graph by keeping a small subset of its arcs. Our goal is to maintain the connectivity required to explain a set of observed traces of information propagation across the graph. Unlike previous work, we do not make any assumption about an underlying model of information propagation. Instead, we approach the task as a combinatorial problem. We prove that the resulting optimization problem is NP -hard. We show that a standard greedy algorithm performs very well in practice, even though it does not have theoretical guarantees. Additionally, if the activity traces have a tree structure, we show that the objective function is supermodular, and experimentally verify that the approach for size-constrained submodular minimization recently proposed by Nagano et al. (28th International Conference on Machine Learning, 2011) produces very good results. Moreover, when applied to the task of reconstructing an unobserved graph, our methods perform comparably to a state-of-the-art algorithm devised specifically for this task.

Original languageEnglish
Pages (from-to)321-343
Number of pages23
JournalData Mining and Knowledge Discovery
Volume27
Issue number3
DOIs
Publication statusPublished - Dec 2013
Externally publishedYes

Fingerprint

Directed graphs
Learning systems
Computational complexity

Keywords

  • Graph sparsification
  • Information propagation
  • Submodular minimization

ASJC Scopus subject areas

  • Information Systems
  • Computer Science Applications
  • Computer Networks and Communications

Cite this

Activity preserving graph simplification. / Bonchi, Francesco; Morales, Gianmarco; Gionis, Aristides; Ukkonen, Antti.

In: Data Mining and Knowledge Discovery, Vol. 27, No. 3, 12.2013, p. 321-343.

Research output: Contribution to journalArticle

Bonchi, Francesco ; Morales, Gianmarco ; Gionis, Aristides ; Ukkonen, Antti. / Activity preserving graph simplification. In: Data Mining and Knowledge Discovery. 2013 ; Vol. 27, No. 3. pp. 321-343.
@article{e93d957c290f4a5282520b2608408db5,
title = "Activity preserving graph simplification",
abstract = "We study the problem of simplifying a given directed graph by keeping a small subset of its arcs. Our goal is to maintain the connectivity required to explain a set of observed traces of information propagation across the graph. Unlike previous work, we do not make any assumption about an underlying model of information propagation. Instead, we approach the task as a combinatorial problem. We prove that the resulting optimization problem is NP -hard. We show that a standard greedy algorithm performs very well in practice, even though it does not have theoretical guarantees. Additionally, if the activity traces have a tree structure, we show that the objective function is supermodular, and experimentally verify that the approach for size-constrained submodular minimization recently proposed by Nagano et al. (28th International Conference on Machine Learning, 2011) produces very good results. Moreover, when applied to the task of reconstructing an unobserved graph, our methods perform comparably to a state-of-the-art algorithm devised specifically for this task.",
keywords = "Graph sparsification, Information propagation, Submodular minimization",
author = "Francesco Bonchi and Gianmarco Morales and Aristides Gionis and Antti Ukkonen",
year = "2013",
month = "12",
doi = "10.1007/s10618-013-0328-8",
language = "English",
volume = "27",
pages = "321--343",
journal = "Data Mining and Knowledge Discovery",
issn = "1384-5810",
publisher = "Springer Netherlands",
number = "3",

}

TY - JOUR

T1 - Activity preserving graph simplification

AU - Bonchi, Francesco

AU - Morales, Gianmarco

AU - Gionis, Aristides

AU - Ukkonen, Antti

PY - 2013/12

Y1 - 2013/12

N2 - We study the problem of simplifying a given directed graph by keeping a small subset of its arcs. Our goal is to maintain the connectivity required to explain a set of observed traces of information propagation across the graph. Unlike previous work, we do not make any assumption about an underlying model of information propagation. Instead, we approach the task as a combinatorial problem. We prove that the resulting optimization problem is NP -hard. We show that a standard greedy algorithm performs very well in practice, even though it does not have theoretical guarantees. Additionally, if the activity traces have a tree structure, we show that the objective function is supermodular, and experimentally verify that the approach for size-constrained submodular minimization recently proposed by Nagano et al. (28th International Conference on Machine Learning, 2011) produces very good results. Moreover, when applied to the task of reconstructing an unobserved graph, our methods perform comparably to a state-of-the-art algorithm devised specifically for this task.

AB - We study the problem of simplifying a given directed graph by keeping a small subset of its arcs. Our goal is to maintain the connectivity required to explain a set of observed traces of information propagation across the graph. Unlike previous work, we do not make any assumption about an underlying model of information propagation. Instead, we approach the task as a combinatorial problem. We prove that the resulting optimization problem is NP -hard. We show that a standard greedy algorithm performs very well in practice, even though it does not have theoretical guarantees. Additionally, if the activity traces have a tree structure, we show that the objective function is supermodular, and experimentally verify that the approach for size-constrained submodular minimization recently proposed by Nagano et al. (28th International Conference on Machine Learning, 2011) produces very good results. Moreover, when applied to the task of reconstructing an unobserved graph, our methods perform comparably to a state-of-the-art algorithm devised specifically for this task.

KW - Graph sparsification

KW - Information propagation

KW - Submodular minimization

UR - http://www.scopus.com/inward/record.url?scp=84881251331&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84881251331&partnerID=8YFLogxK

U2 - 10.1007/s10618-013-0328-8

DO - 10.1007/s10618-013-0328-8

M3 - Article

AN - SCOPUS:84881251331

VL - 27

SP - 321

EP - 343

JO - Data Mining and Knowledge Discovery

JF - Data Mining and Knowledge Discovery

SN - 1384-5810

IS - 3

ER -