A new approach to exploring architecture of bipartite (interaction) ecological networks

J. Podaní, Ferenc Jordán, D. Schmera

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

We propose the use of a new methodological scheme for exploring and quantifying structure in bipartite ecological networks. In this, graphical visualization and numerical measurement are combined, offering a unique possibility for network analysis on a coherent conceptual basis. Dissimilarity between all species pairs in either group of constituting species, calculated based on their interactions with the other group, is decomposed into additive fractions expressing link replacement and degree difference of nodes. Together with interaction similarity, these are used to visualize network features in form of a two-dimensional simplex diagram. Furthermore, these quantities serve as a basis for characterizing and measuring nestedness, architectural asymmetry as well as bimodality in the distribution of interaction dissimilarity, which in turn can be used to detect modularity in the graph. The approach is extended to weighted links, facilitating comparisons with the unweighted version of the same network, which was an old but unresolved issue in network science. This is especially important for nestedness and modularity whose measurement has been confined mostly to unweighted networks. The use of the method is demonstrated by artificial graphs and on selected examples of actual ecological network data. We suggest that the methods apply equally well to bipartite graphs used in other fields of science as well.

Original languageEnglish
Pages (from-to)168-186
Number of pages19
JournalJournal of Complex Networks
Volume2
Issue number2
DOIs
Publication statusPublished - 2014

Fingerprint

Electric network analysis
Interaction
Visualization
Dissimilarity
Modularity
Bimodality
Network Analysis
Graph in graph theory
Bipartite Graph
Asymmetry
Replacement
Architecture
Diagram
Vertex of a graph
Graph

Keywords

  • Asymmetry
  • Connectance
  • Jaccard dissimilarity
  • Modularity
  • Nestedness
  • Simplex

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Management Science and Operations Research
  • Applied Mathematics
  • Computational Mathematics
  • Control and Optimization

Cite this

A new approach to exploring architecture of bipartite (interaction) ecological networks. / Podaní, J.; Jordán, Ferenc; Schmera, D.

In: Journal of Complex Networks, Vol. 2, No. 2, 2014, p. 168-186.

Research output: Contribution to journalArticle

@article{997f1ec0c0d846ec836f764e05f3891f,
title = "A new approach to exploring architecture of bipartite (interaction) ecological networks",
abstract = "We propose the use of a new methodological scheme for exploring and quantifying structure in bipartite ecological networks. In this, graphical visualization and numerical measurement are combined, offering a unique possibility for network analysis on a coherent conceptual basis. Dissimilarity between all species pairs in either group of constituting species, calculated based on their interactions with the other group, is decomposed into additive fractions expressing link replacement and degree difference of nodes. Together with interaction similarity, these are used to visualize network features in form of a two-dimensional simplex diagram. Furthermore, these quantities serve as a basis for characterizing and measuring nestedness, architectural asymmetry as well as bimodality in the distribution of interaction dissimilarity, which in turn can be used to detect modularity in the graph. The approach is extended to weighted links, facilitating comparisons with the unweighted version of the same network, which was an old but unresolved issue in network science. This is especially important for nestedness and modularity whose measurement has been confined mostly to unweighted networks. The use of the method is demonstrated by artificial graphs and on selected examples of actual ecological network data. We suggest that the methods apply equally well to bipartite graphs used in other fields of science as well.",
keywords = "Asymmetry, Connectance, Jaccard dissimilarity, Modularity, Nestedness, Simplex",
author = "J. Podan{\'i} and Ferenc Jord{\'a}n and D. Schmera",
year = "2014",
doi = "10.1093/comnet/cnu002",
language = "English",
volume = "2",
pages = "168--186",
journal = "Journal of Complex Networks",
issn = "2051-1310",
publisher = "Oxford University Press",
number = "2",

}

TY - JOUR

T1 - A new approach to exploring architecture of bipartite (interaction) ecological networks

AU - Podaní, J.

AU - Jordán, Ferenc

AU - Schmera, D.

PY - 2014

Y1 - 2014

N2 - We propose the use of a new methodological scheme for exploring and quantifying structure in bipartite ecological networks. In this, graphical visualization and numerical measurement are combined, offering a unique possibility for network analysis on a coherent conceptual basis. Dissimilarity between all species pairs in either group of constituting species, calculated based on their interactions with the other group, is decomposed into additive fractions expressing link replacement and degree difference of nodes. Together with interaction similarity, these are used to visualize network features in form of a two-dimensional simplex diagram. Furthermore, these quantities serve as a basis for characterizing and measuring nestedness, architectural asymmetry as well as bimodality in the distribution of interaction dissimilarity, which in turn can be used to detect modularity in the graph. The approach is extended to weighted links, facilitating comparisons with the unweighted version of the same network, which was an old but unresolved issue in network science. This is especially important for nestedness and modularity whose measurement has been confined mostly to unweighted networks. The use of the method is demonstrated by artificial graphs and on selected examples of actual ecological network data. We suggest that the methods apply equally well to bipartite graphs used in other fields of science as well.

AB - We propose the use of a new methodological scheme for exploring and quantifying structure in bipartite ecological networks. In this, graphical visualization and numerical measurement are combined, offering a unique possibility for network analysis on a coherent conceptual basis. Dissimilarity between all species pairs in either group of constituting species, calculated based on their interactions with the other group, is decomposed into additive fractions expressing link replacement and degree difference of nodes. Together with interaction similarity, these are used to visualize network features in form of a two-dimensional simplex diagram. Furthermore, these quantities serve as a basis for characterizing and measuring nestedness, architectural asymmetry as well as bimodality in the distribution of interaction dissimilarity, which in turn can be used to detect modularity in the graph. The approach is extended to weighted links, facilitating comparisons with the unweighted version of the same network, which was an old but unresolved issue in network science. This is especially important for nestedness and modularity whose measurement has been confined mostly to unweighted networks. The use of the method is demonstrated by artificial graphs and on selected examples of actual ecological network data. We suggest that the methods apply equally well to bipartite graphs used in other fields of science as well.

KW - Asymmetry

KW - Connectance

KW - Jaccard dissimilarity

KW - Modularity

KW - Nestedness

KW - Simplex

UR - http://www.scopus.com/inward/record.url?scp=84923350404&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84923350404&partnerID=8YFLogxK

U2 - 10.1093/comnet/cnu002

DO - 10.1093/comnet/cnu002

M3 - Article

AN - SCOPUS:84923350404

VL - 2

SP - 168

EP - 186

JO - Journal of Complex Networks

JF - Journal of Complex Networks

SN - 2051-1310

IS - 2

ER -