Automatic discovery of global and local equivalence relationships in labeled geo-spatial data

Bart Thomee, Gianmarco Morales

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

We propose a novel algorithmic framework to automatically detect which labels refer to the same concept in labeled spatial data. People often use different words and synonyms when referring to the same concept or location. Furthermore these words and their usage vary across culture, language, and place. Our method analyzes the patterns in the spatial distribution of labels to discover equivalence relationships. We evaluate our proposed technique on a large collection of geo-referenced Flickr photos using a semi-automatically constructed ground truth from an existing ontology. Our approach is able to classify equivalent tags with a high accuracy (AUC of 0.85), as well as providing the geographic extent where the relationship holds.

Original languageEnglish
Title of host publicationHT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media
PublisherAssociation for Computing Machinery
Pages158-168
Number of pages11
ISBN (Print)9781450329545
DOIs
Publication statusPublished - 2014
Externally publishedYes
Event25th ACM Conference on Hypertext and Social Media, HT 2014 - Santiago, Chile
Duration: 1 Sep 20144 Sep 2014

Other

Other25th ACM Conference on Hypertext and Social Media, HT 2014
CountryChile
CitySantiago
Period1/9/144/9/14

Fingerprint

Labels
Spatial distribution
Ontology

Keywords

  • flickr
  • folksonomy
  • geo-spatial analysis
  • geotagged data
  • relationship discovery

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Human-Computer Interaction
  • Software

Cite this

Thomee, B., & Morales, G. (2014). Automatic discovery of global and local equivalence relationships in labeled geo-spatial data. In HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media (pp. 158-168). Association for Computing Machinery. https://doi.org/10.1145/2631775.2631794

Automatic discovery of global and local equivalence relationships in labeled geo-spatial data. / Thomee, Bart; Morales, Gianmarco.

HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media. Association for Computing Machinery, 2014. p. 158-168.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Thomee, B & Morales, G 2014, Automatic discovery of global and local equivalence relationships in labeled geo-spatial data. in HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media. Association for Computing Machinery, pp. 158-168, 25th ACM Conference on Hypertext and Social Media, HT 2014, Santiago, Chile, 1/9/14. https://doi.org/10.1145/2631775.2631794
Thomee B, Morales G. Automatic discovery of global and local equivalence relationships in labeled geo-spatial data. In HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media. Association for Computing Machinery. 2014. p. 158-168 https://doi.org/10.1145/2631775.2631794
Thomee, Bart ; Morales, Gianmarco. / Automatic discovery of global and local equivalence relationships in labeled geo-spatial data. HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media. Association for Computing Machinery, 2014. pp. 158-168
@inproceedings{ef53924f8cdf455fbfd67812482d8e42,
title = "Automatic discovery of global and local equivalence relationships in labeled geo-spatial data",
abstract = "We propose a novel algorithmic framework to automatically detect which labels refer to the same concept in labeled spatial data. People often use different words and synonyms when referring to the same concept or location. Furthermore these words and their usage vary across culture, language, and place. Our method analyzes the patterns in the spatial distribution of labels to discover equivalence relationships. We evaluate our proposed technique on a large collection of geo-referenced Flickr photos using a semi-automatically constructed ground truth from an existing ontology. Our approach is able to classify equivalent tags with a high accuracy (AUC of 0.85), as well as providing the geographic extent where the relationship holds.",
keywords = "flickr, folksonomy, geo-spatial analysis, geotagged data, relationship discovery",
author = "Bart Thomee and Gianmarco Morales",
year = "2014",
doi = "10.1145/2631775.2631794",
language = "English",
isbn = "9781450329545",
pages = "158--168",
booktitle = "HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Automatic discovery of global and local equivalence relationships in labeled geo-spatial data

AU - Thomee, Bart

AU - Morales, Gianmarco

PY - 2014

Y1 - 2014

N2 - We propose a novel algorithmic framework to automatically detect which labels refer to the same concept in labeled spatial data. People often use different words and synonyms when referring to the same concept or location. Furthermore these words and their usage vary across culture, language, and place. Our method analyzes the patterns in the spatial distribution of labels to discover equivalence relationships. We evaluate our proposed technique on a large collection of geo-referenced Flickr photos using a semi-automatically constructed ground truth from an existing ontology. Our approach is able to classify equivalent tags with a high accuracy (AUC of 0.85), as well as providing the geographic extent where the relationship holds.

AB - We propose a novel algorithmic framework to automatically detect which labels refer to the same concept in labeled spatial data. People often use different words and synonyms when referring to the same concept or location. Furthermore these words and their usage vary across culture, language, and place. Our method analyzes the patterns in the spatial distribution of labels to discover equivalence relationships. We evaluate our proposed technique on a large collection of geo-referenced Flickr photos using a semi-automatically constructed ground truth from an existing ontology. Our approach is able to classify equivalent tags with a high accuracy (AUC of 0.85), as well as providing the geographic extent where the relationship holds.

KW - flickr

KW - folksonomy

KW - geo-spatial analysis

KW - geotagged data

KW - relationship discovery

UR - http://www.scopus.com/inward/record.url?scp=84907405317&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907405317&partnerID=8YFLogxK

U2 - 10.1145/2631775.2631794

DO - 10.1145/2631775.2631794

M3 - Conference contribution

SN - 9781450329545

SP - 158

EP - 168

BT - HT 2014 - Proceedings of the 25th ACM Conference on Hypertext and Social Media

PB - Association for Computing Machinery

ER -