Change search
ReferencesLink to record
Permanent link

Direct link
Exploring geographical metadata by automatic and visual data mining
KTH, Superseded Departments, Infrastructure.
2004 (English)Licentiate thesis, monograph (Other scientific)
Abstract [en]

Metadata are data about data. They describe characteristicsand content of an original piece of data. Geographical metadatadescribe geospatial data: maps, satellite images and othergeographically referenced material. Such metadata have twocharacteristics, high dimensionality and diversity of attributedata types, which present a problem for traditional data miningalgorithms.

Other problems that arise during the exploration ofgeographical metadata are linked to the expertise of the userperforming the analysis. The large amounts of metadata andhundreds of possible attributes limit the exploration for anon-expert user, which results in a potential loss ofinformation that is hidden in metadata.

In order to solve some of these problems, this thesispresents an approach for exploration of geographical metadataby a combination of automatic and visual data mining.

Visual data mining is a principle that involves the human inthe data exploration by presenting the data in some visualform, allowing the human to get insight into the data and torecognise patterns. The main advantages of visual dataexploration over automatic data mining are that the visualexploration allows a direct interaction with the user, that itis intuitive and does not require complex understanding ofmathematical or statistical algorithms. As a result the userhas a higher confidence in the resulting patterns than if theywere produced by computer only.

In the thesis we present the Visual data mining tool (VDMtool), which was developed for exploration of geographicalmetadata for site planning. The tool provides five differentvisualisations: a histogram, a table, a pie chart, a parallelcoordinates visualisation and a clustering visualisation. Thevisualisations are connected using the interactive selectionprinciple called brushing and linking.

In the VDM tool the visual data mining concept is integratedwith an automatic data mining method, clustering, which finds ahierarchical structure in the metadata, based on similarity ofmetadata items. In the thesis we present a visualisation of thehierarchical structure in the form of a snowflake graph.

Keywords:visualisation, data mining, clustering, treedrawing, geographical metadata.

Place, publisher, year, edition, pages
Stockholm: Infrastruktur , 2004. , vii, 93 p.
Trita-INFRA, ISSN 1651-0216 ; 04:010
URN: urn:nbn:se:kth:diva-1779ISBN: 91-7323-077-4OAI: diva2:7771
Available from: 2004-07-06 Created: 2004-07-06 Last updated: 2012-03-20

Open Access in DiVA

fulltext(7152 kB)1618 downloads
File information
File name FULLTEXT01.pdfFile size 7152 kBChecksum SHA-1
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Demšar, Urška
By organisation

Search outside of DiVA

GoogleGoogle Scholar
Total: 1618 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 828 hits
ReferencesLink to record
Permanent link

Direct link