Design and evaluation of a layered thematic knowledge map system

Sheng Tun Li, Won Chen Chang

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Knowledge maps (K-maps) have been recognized as an expressive, efficient and effective way to enhance access and navigation of information in a large knowledge repository. However, its capability is limited by the lack of a way to systematically discover the hidden structure of codified knowledge and compactly visualize the map morphology to improve effective navigation. This paper presents a new hybrid approach to tackle these issues and proposes a layered thematic K-map system. The knowledge objects are firstly categorized and labeled according to their hidden conceptual architecture using a hierarchical self-organizing map network and then thematically navigated through a spatially expandable stacked view with corresponding information. The system is tested with a real-world soil remediation patent corpus with various map topologies. A pair-wise evaluation consisting of clustering performance and usability assessment was conducted. Compared with a typical file/directory view, the results of usability evaluation showed that the proposed system enhances the efficiency of comparison tasks while preserving comparable performance of identification and association tasks.

Original languageEnglish
Pages (from-to)92-103
Number of pages12
JournalJournal of Computer Information Systems
Volume49
Issue number2
Publication statusPublished - 2008 Dec 1

Fingerprint

evaluation
Navigation
patent
performance
Self organizing maps
Remediation
efficiency
lack
Topology
Soils
knowledge

All Science Journal Classification (ASJC) codes

  • Information Systems
  • Education
  • Computer Networks and Communications

Cite this

@article{e98f8360f85c4d299f63db7b08f46c43,
title = "Design and evaluation of a layered thematic knowledge map system",
abstract = "Knowledge maps (K-maps) have been recognized as an expressive, efficient and effective way to enhance access and navigation of information in a large knowledge repository. However, its capability is limited by the lack of a way to systematically discover the hidden structure of codified knowledge and compactly visualize the map morphology to improve effective navigation. This paper presents a new hybrid approach to tackle these issues and proposes a layered thematic K-map system. The knowledge objects are firstly categorized and labeled according to their hidden conceptual architecture using a hierarchical self-organizing map network and then thematically navigated through a spatially expandable stacked view with corresponding information. The system is tested with a real-world soil remediation patent corpus with various map topologies. A pair-wise evaluation consisting of clustering performance and usability assessment was conducted. Compared with a typical file/directory view, the results of usability evaluation showed that the proposed system enhances the efficiency of comparison tasks while preserving comparable performance of identification and association tasks.",
author = "Li, {Sheng Tun} and Chang, {Won Chen}",
year = "2008",
month = "12",
day = "1",
language = "English",
volume = "49",
pages = "92--103",
journal = "Journal of Computer Information Systems",
issn = "0887-4417",
publisher = "International Association for Computer Information Systems",
number = "2",

}

Design and evaluation of a layered thematic knowledge map system. / Li, Sheng Tun; Chang, Won Chen.

In: Journal of Computer Information Systems, Vol. 49, No. 2, 01.12.2008, p. 92-103.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Design and evaluation of a layered thematic knowledge map system

AU - Li, Sheng Tun

AU - Chang, Won Chen

PY - 2008/12/1

Y1 - 2008/12/1

N2 - Knowledge maps (K-maps) have been recognized as an expressive, efficient and effective way to enhance access and navigation of information in a large knowledge repository. However, its capability is limited by the lack of a way to systematically discover the hidden structure of codified knowledge and compactly visualize the map morphology to improve effective navigation. This paper presents a new hybrid approach to tackle these issues and proposes a layered thematic K-map system. The knowledge objects are firstly categorized and labeled according to their hidden conceptual architecture using a hierarchical self-organizing map network and then thematically navigated through a spatially expandable stacked view with corresponding information. The system is tested with a real-world soil remediation patent corpus with various map topologies. A pair-wise evaluation consisting of clustering performance and usability assessment was conducted. Compared with a typical file/directory view, the results of usability evaluation showed that the proposed system enhances the efficiency of comparison tasks while preserving comparable performance of identification and association tasks.

AB - Knowledge maps (K-maps) have been recognized as an expressive, efficient and effective way to enhance access and navigation of information in a large knowledge repository. However, its capability is limited by the lack of a way to systematically discover the hidden structure of codified knowledge and compactly visualize the map morphology to improve effective navigation. This paper presents a new hybrid approach to tackle these issues and proposes a layered thematic K-map system. The knowledge objects are firstly categorized and labeled according to their hidden conceptual architecture using a hierarchical self-organizing map network and then thematically navigated through a spatially expandable stacked view with corresponding information. The system is tested with a real-world soil remediation patent corpus with various map topologies. A pair-wise evaluation consisting of clustering performance and usability assessment was conducted. Compared with a typical file/directory view, the results of usability evaluation showed that the proposed system enhances the efficiency of comparison tasks while preserving comparable performance of identification and association tasks.

UR - http://www.scopus.com/inward/record.url?scp=60849111681&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=60849111681&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:60849111681

VL - 49

SP - 92

EP - 103

JO - Journal of Computer Information Systems

JF - Journal of Computer Information Systems

SN - 0887-4417

IS - 2

ER -