Cluster validity and uncertainty assessment for self-organizing map pest profile analysis

Mariona Roigé, Melodie A. McGeoch, Cang Hui, Susan P. Worner

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Pest risk assessment (PRA) comprises a set of quantitative and qualitative tools to protect productive ecosystems from the impacts of unwanted biological invasions. Self-organizing maps for pest profile analysis (SOM PPA) is a methodological approach aimed to support PRA. It is based on cluster analysis and extracts information out of current distributions of insect crop pests world-wide, allowing the analyst to generate a list of potential risk species for a target region. Self-organizing maps for pest profile analysis currently lacks of a measure of performance able to provide a level of confidence for its outputs. In this study, we investigate ζ diversity as an ecologically meaningful and generalizable metric of similarity. The application of ζ allowed us to quantify and thus reveal different levels of similarity across pest profiles. The use of ζ diversity applied to the SOM PPA provides an informative measure of uncertainty for the output of SOM PPA, thus adding major improvements to the methodology while only marginally increasing its complexity.

Original languageEnglish
Pages (from-to)349-357
Number of pages9
JournalMethods in Ecology and Evolution
Volume8
Issue number3
DOIs
Publication statusPublished - 1 Mar 2017

Keywords

  • cluster validity
  • pest profile analysis
  • pest risk assessment
  • self-organizing maps
  • uncertainty
  • zeta diversity

Cite this

@article{64fb64aa4d4d41b78e731d3e1979aa1b,
title = "Cluster validity and uncertainty assessment for self-organizing map pest profile analysis",
abstract = "Pest risk assessment (PRA) comprises a set of quantitative and qualitative tools to protect productive ecosystems from the impacts of unwanted biological invasions. Self-organizing maps for pest profile analysis (SOM PPA) is a methodological approach aimed to support PRA. It is based on cluster analysis and extracts information out of current distributions of insect crop pests world-wide, allowing the analyst to generate a list of potential risk species for a target region. Self-organizing maps for pest profile analysis currently lacks of a measure of performance able to provide a level of confidence for its outputs. In this study, we investigate ζ diversity as an ecologically meaningful and generalizable metric of similarity. The application of ζ allowed us to quantify and thus reveal different levels of similarity across pest profiles. The use of ζ diversity applied to the SOM PPA provides an informative measure of uncertainty for the output of SOM PPA, thus adding major improvements to the methodology while only marginally increasing its complexity.",
keywords = "cluster validity, pest profile analysis, pest risk assessment, self-organizing maps, uncertainty, zeta diversity",
author = "Mariona Roig{\'e} and McGeoch, {Melodie A.} and Cang Hui and Worner, {Susan P.}",
year = "2017",
month = "3",
day = "1",
doi = "10.1111/2041-210X.12669",
language = "English",
volume = "8",
pages = "349--357",
journal = "Methods in Ecology and Evolution",
issn = "2041-210X",
publisher = "Wiley-Blackwell",
number = "3",

}

Cluster validity and uncertainty assessment for self-organizing map pest profile analysis. / Roigé, Mariona; McGeoch, Melodie A.; Hui, Cang; Worner, Susan P.

In: Methods in Ecology and Evolution, Vol. 8, No. 3, 01.03.2017, p. 349-357.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Cluster validity and uncertainty assessment for self-organizing map pest profile analysis

AU - Roigé, Mariona

AU - McGeoch, Melodie A.

AU - Hui, Cang

AU - Worner, Susan P.

PY - 2017/3/1

Y1 - 2017/3/1

N2 - Pest risk assessment (PRA) comprises a set of quantitative and qualitative tools to protect productive ecosystems from the impacts of unwanted biological invasions. Self-organizing maps for pest profile analysis (SOM PPA) is a methodological approach aimed to support PRA. It is based on cluster analysis and extracts information out of current distributions of insect crop pests world-wide, allowing the analyst to generate a list of potential risk species for a target region. Self-organizing maps for pest profile analysis currently lacks of a measure of performance able to provide a level of confidence for its outputs. In this study, we investigate ζ diversity as an ecologically meaningful and generalizable metric of similarity. The application of ζ allowed us to quantify and thus reveal different levels of similarity across pest profiles. The use of ζ diversity applied to the SOM PPA provides an informative measure of uncertainty for the output of SOM PPA, thus adding major improvements to the methodology while only marginally increasing its complexity.

AB - Pest risk assessment (PRA) comprises a set of quantitative and qualitative tools to protect productive ecosystems from the impacts of unwanted biological invasions. Self-organizing maps for pest profile analysis (SOM PPA) is a methodological approach aimed to support PRA. It is based on cluster analysis and extracts information out of current distributions of insect crop pests world-wide, allowing the analyst to generate a list of potential risk species for a target region. Self-organizing maps for pest profile analysis currently lacks of a measure of performance able to provide a level of confidence for its outputs. In this study, we investigate ζ diversity as an ecologically meaningful and generalizable metric of similarity. The application of ζ allowed us to quantify and thus reveal different levels of similarity across pest profiles. The use of ζ diversity applied to the SOM PPA provides an informative measure of uncertainty for the output of SOM PPA, thus adding major improvements to the methodology while only marginally increasing its complexity.

KW - cluster validity

KW - pest profile analysis

KW - pest risk assessment

KW - self-organizing maps

KW - uncertainty

KW - zeta diversity

UR - http://www.scopus.com/inward/record.url?scp=84995592221&partnerID=8YFLogxK

U2 - 10.1111/2041-210X.12669

DO - 10.1111/2041-210X.12669

M3 - Article

VL - 8

SP - 349

EP - 357

JO - Methods in Ecology and Evolution

JF - Methods in Ecology and Evolution

SN - 2041-210X

IS - 3

ER -