Show simple item record

dc.contributor.authorMoghaddam, Mehdi Minachien
dc.date.accessioned2014-03-14T10:10:52Z
dc.date.available2014-03-14T10:10:52Z
dc.date.issued2005-01
dc.identifier.citationMoghaddam, M.M. (2005) 'Internet search techniques: using word count, links and directory structure as internet search tools'. PhD thesis. University of Luton.en
dc.identifier.urihttp://hdl.handle.net/10547/314080
dc.descriptionA thesis submitted for the degree of Doctor of Philosophy ofthe University of Lutonen
dc.description.abstractAs the Web grows in size it becomes increasingly important that ways are developed to maximise the efficiency of the search process and index its contents with minimal human intervention. An evaluation is undertaken of current popular search engines which use a centralised index approach. Using a number of search terms and metrics that measure similarity between sets of results, it was found that there is very little commonality between the outcome of the same search performed using different search engines. A semi-automated system for searching the web is presented, the Internet Search Agent (ISA), this employs a method for indexing based upon the idea of "fingerprint types". These fingerprint types are based upon the text and links contained in the web pages being indexed. Three examples of fingerprint type are developed, the first concentrating upon the textual content of the indexed files, the other two augment this with the use of links to and from these files. By looking at the results returned as a search progresses in terms of numbers and measures of content of results for effort expended, comparisons can be made between the three fingerprint types. The ISA model allows the searcher to be presented with results in context and potentially allows for distributed searching to be implemented.
dc.language.isoenen
dc.publisherUniversity of Bedfordshireen
dc.subjectG440 Human-computer Interactionen
dc.subjectinternet searchen
dc.subjectsearchen
dc.subjectsearch toolsen
dc.titleInternet search techniques: using word count, links and directory structure as internet search toolsen
dc.typeThesis or dissertationen
dc.type.qualificationnamePhDen_GB
dc.type.qualificationlevelPhDen
dc.publisher.institutionUniversity of Bedfordshireen
refterms.dateFOA2020-05-12T13:02:02Z
html.description.abstractAs the Web grows in size it becomes increasingly important that ways are developed to maximise the efficiency of the search process and index its contents with minimal human intervention. An evaluation is undertaken of current popular search engines which use a centralised index approach. Using a number of search terms and metrics that measure similarity between sets of results, it was found that there is very little commonality between the outcome of the same search performed using different search engines. A semi-automated system for searching the web is presented, the Internet Search Agent (ISA), this employs a method for indexing based upon the idea of "fingerprint types". These fingerprint types are based upon the text and links contained in the web pages being indexed. Three examples of fingerprint type are developed, the first concentrating upon the textual content of the indexed files, the other two augment this with the use of links to and from these files. By looking at the results returned as a search progresses in terms of numbers and measures of content of results for effort expended, comparisons can be made between the three fingerprint types. The ISA model allows the searcher to be presented with results in context and potentially allows for distributed searching to be implemented.


Files in this item

Thumbnail
Name:
M Minachin Moghaddam 1.pdf
Size:
40.62Mb
Format:
PDF
Description:
Thesis

This item appears in the following Collection(s)

Show simple item record