Show simple item record

dc.contributor.authorThelwall, Mike
dc.date.accessioned2006-08-24T10:05:54Z
dc.date.available2006-08-24T10:05:54Z
dc.date.issued2002
dc.identifier.citationAslib Proceedings, 54(2): 118-128
dc.identifier.issn0001253X
dc.identifier.doi10.1108/00012530210435248
dc.identifier.urihttp://hdl.handle.net/2436/4016
dc.description.abstractAggregates of links are of interest to information scientists in the same way as citation counts are: as potential sources of data from which new knowledge can be mined. Builds on the recent discovery of a correlation between a Web link count measure and the research quality of British universities by applying a range of multivariate statistical techniques to counts of links between pairs of universities. This represents an initial attempt at developing an understanding of this phenomenon. Extracts plausible results. Also identifies outliers in the data by the techniques, some of which were verified by being tracked down to identifiable Web phenomena. This is an important outcome because successful anomaly identification is a precondition to more effective analysis of this kind of data. The identification of groupings is encouraging evidence that Web links between universities can be mined for significant results, although it is clear that more methodological development is needed, if any but the simplest patterns are to be extracted. Finally, based upon the types of patterns extracted, argues that none of the methods used are capable of fully analysing link structures on their own.
dc.format.extent239880 bytes
dc.format.mimetypeapplication/pdf
dc.language.isoen
dc.publisherMCB UP Ltd
dc.relation.urlhttp://www.emeraldinsight.com/10.1108/00012530210435248
dc.subjectWeblinks
dc.subjectAcademic websites
dc.subjectInternet
dc.subjectUniversities
dc.subjectUser studies
dc.subjectNetworks
dc.titleAn initial exploration of the link relationship between UK university Web sites.
dc.typeJournal article
dc.format.digYES
refterms.dateFOA2018-08-20T16:12:21Z
html.description.abstractAggregates of links are of interest to information scientists in the same way as citation counts are: as potential sources of data from which new knowledge can be mined. Builds on the recent discovery of a correlation between a Web link count measure and the research quality of British universities by applying a range of multivariate statistical techniques to counts of links between pairs of universities. This represents an initial attempt at developing an understanding of this phenomenon. Extracts plausible results. Also identifies outliers in the data by the techniques, some of which were verified by being tracked down to identifiable Web phenomena. This is an important outcome because successful anomaly identification is a precondition to more effective analysis of this kind of data. The identification of groupings is encouraging evidence that Web links between universities can be mined for significant results, although it is clear that more methodological development is needed, if any but the simplest patterns are to be extracted. Finally, based upon the types of patterns extracted, argues that none of the methods used are capable of fully analysing link structures on their own.


Files in this item

Thumbnail
Name:
2002_An_initial_exploration_of ...
Size:
234.2Kb
Format:
PDF

This item appears in the following Collection(s)

Show simple item record