Extracting Attractive Local-Area Topics in Georeferenced Documents using a New Density-based Spatial Clustering Algorithm

Along with the popularization of social media, huge numbers of georeferenced documents (which include location information) are being posted on social media sites via the Internet, allowing people to transmit and collect geographic information. Typically, such georeferenced documents are related not...

Full description

Saved in:
Bibliographic Details
Published in:IAENG international journal of computer science Vol. 41; no. 3; pp. 185 - 192
Main Authors: Sakai, Tatsuhiro, Tamura, Keiichi, Kitakami, Hajime
Format: Journal Article
Language:English
Published: 01.09.2014
Subjects:
Algorithms
Clustering
Clusters
Digital media
Internet
Recognition
Social networks
ISSN:1819-656X, 1819-9224
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Along with the popularization of social media, huge numbers of georeferenced documents (which include location information) are being posted on social media sites via the Internet, allowing people to transmit and collect geographic information. Typically, such georeferenced documents are related not only to personal topics but also to local topics and events. Therefore, extracting "attractive" areas associated with local topics from georeferenced documents is currently one of the most important challenges in different application domains. In this paper, a novel spatial clustering algorithm for extracting "attractive" local-area topics in georeferenced documents, known as the ([epsilon], [sigma])-density-based spatial clustering algorithm, is proposed. We defined a new type of spatial cluster called an ([epsilon], [sigma])-density-based spatial cluster. The proposed density-based spatial clustering algorithm can recognize both semantically and spatially separated spatial clusters. Therefore, the proposed algorithm can extract "attractive" local-area topics as ([epsilon], [sigma])-density-based spatial clusters. To evaluate our proposed clustering algorithm, geo-tagged tweets posted on the Twitter site were used. The experimental results showed that the ([epsilon], [sigma])-density-based spatial clustering algorithm could extract "attractive" areas as the ([epsilon], [sigma])-density-based spatial clusters that were closely related to local topics.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:1819-656X
1819-9224