Predicting potential distributions of geographic events using one-class data: Concepts and methods

Q. Guo, W. Li, Y. Liu, Daoqin Tong

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

One common problem with geographic data is that, for a specific geographic event, only occurrence information is available; information about the absence of the event is not available.We refer to these specific types of geospatial data as geographic one-class data (GOCD). Predicting the potential spatial distributions that a particular geographic event may occur from GOCD is difficult because traditional binary classification methods that require availability of both positive and negative training samples cannot be used. The objective of this research is to define GOCD and propose novel approaches for modelling potential spatial distributions of geographic events using GOCD. We investigate the effectiveness of one-class support vector machine (OCSVM), maximum entropy (MAXENT) and the newly proposed positive and unlabelled learning (PUL) algorithm for solving GOCD problems using a case study: species distribution modelling from synthetic data. Our experimental results indicate that generally OCSVM, MAXENT and PUL are effective in modelling the GOCD. Each method has advantages and disadvantages, but PUL seems to be the most promising method.

Original languageEnglish (US)
Pages (from-to)1697-1715
Number of pages19
JournalInternational Journal of Geographical Information Science
Volume25
Issue number10
DOIs
StatePublished - Oct 1 2011
Externally publishedYes

Fingerprint

Spatial distribution
Support vector machines
Entropy
event
Learning algorithms
Availability
learning
entropy
spatial distribution
modeling
distribution
method
available information

Keywords

  • Ecological niche modelling
  • Geographic one-class data
  • Maximum entropy
  • One-class support vector machine
  • Positive and unlabelled learning

ASJC Scopus subject areas

  • Information Systems
  • Geography, Planning and Development
  • Library and Information Sciences

Cite this

Predicting potential distributions of geographic events using one-class data : Concepts and methods. / Guo, Q.; Li, W.; Liu, Y.; Tong, Daoqin.

In: International Journal of Geographical Information Science, Vol. 25, No. 10, 01.10.2011, p. 1697-1715.

Research output: Contribution to journalArticle

@article{bf90f81bd4284b8888207c7440f05ff7,
title = "Predicting potential distributions of geographic events using one-class data: Concepts and methods",
abstract = "One common problem with geographic data is that, for a specific geographic event, only occurrence information is available; information about the absence of the event is not available.We refer to these specific types of geospatial data as geographic one-class data (GOCD). Predicting the potential spatial distributions that a particular geographic event may occur from GOCD is difficult because traditional binary classification methods that require availability of both positive and negative training samples cannot be used. The objective of this research is to define GOCD and propose novel approaches for modelling potential spatial distributions of geographic events using GOCD. We investigate the effectiveness of one-class support vector machine (OCSVM), maximum entropy (MAXENT) and the newly proposed positive and unlabelled learning (PUL) algorithm for solving GOCD problems using a case study: species distribution modelling from synthetic data. Our experimental results indicate that generally OCSVM, MAXENT and PUL are effective in modelling the GOCD. Each method has advantages and disadvantages, but PUL seems to be the most promising method.",
keywords = "Ecological niche modelling, Geographic one-class data, Maximum entropy, One-class support vector machine, Positive and unlabelled learning",
author = "Q. Guo and W. Li and Y. Liu and Daoqin Tong",
year = "2011",
month = "10",
day = "1",
doi = "10.1080/13658816.2010.546360",
language = "English (US)",
volume = "25",
pages = "1697--1715",
journal = "International Journal of Geographical Information Science",
issn = "1365-8816",
publisher = "Taylor and Francis Ltd.",
number = "10",

}

TY - JOUR

T1 - Predicting potential distributions of geographic events using one-class data

T2 - Concepts and methods

AU - Guo, Q.

AU - Li, W.

AU - Liu, Y.

AU - Tong, Daoqin

PY - 2011/10/1

Y1 - 2011/10/1

N2 - One common problem with geographic data is that, for a specific geographic event, only occurrence information is available; information about the absence of the event is not available.We refer to these specific types of geospatial data as geographic one-class data (GOCD). Predicting the potential spatial distributions that a particular geographic event may occur from GOCD is difficult because traditional binary classification methods that require availability of both positive and negative training samples cannot be used. The objective of this research is to define GOCD and propose novel approaches for modelling potential spatial distributions of geographic events using GOCD. We investigate the effectiveness of one-class support vector machine (OCSVM), maximum entropy (MAXENT) and the newly proposed positive and unlabelled learning (PUL) algorithm for solving GOCD problems using a case study: species distribution modelling from synthetic data. Our experimental results indicate that generally OCSVM, MAXENT and PUL are effective in modelling the GOCD. Each method has advantages and disadvantages, but PUL seems to be the most promising method.

AB - One common problem with geographic data is that, for a specific geographic event, only occurrence information is available; information about the absence of the event is not available.We refer to these specific types of geospatial data as geographic one-class data (GOCD). Predicting the potential spatial distributions that a particular geographic event may occur from GOCD is difficult because traditional binary classification methods that require availability of both positive and negative training samples cannot be used. The objective of this research is to define GOCD and propose novel approaches for modelling potential spatial distributions of geographic events using GOCD. We investigate the effectiveness of one-class support vector machine (OCSVM), maximum entropy (MAXENT) and the newly proposed positive and unlabelled learning (PUL) algorithm for solving GOCD problems using a case study: species distribution modelling from synthetic data. Our experimental results indicate that generally OCSVM, MAXENT and PUL are effective in modelling the GOCD. Each method has advantages and disadvantages, but PUL seems to be the most promising method.

KW - Ecological niche modelling

KW - Geographic one-class data

KW - Maximum entropy

KW - One-class support vector machine

KW - Positive and unlabelled learning

UR - http://www.scopus.com/inward/record.url?scp=84862908217&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84862908217&partnerID=8YFLogxK

U2 - 10.1080/13658816.2010.546360

DO - 10.1080/13658816.2010.546360

M3 - Article

AN - SCOPUS:84862908217

VL - 25

SP - 1697

EP - 1715

JO - International Journal of Geographical Information Science

JF - International Journal of Geographical Information Science

SN - 1365-8816

IS - 10

ER -