dc.contributor.advisor | Midtstraum, Roger | nb_NO |
dc.contributor.advisor | Zoric, Josip | nb_NO |
dc.contributor.author | Kismul, Hans Kristian | nb_NO |
dc.date.accessioned | 2014-12-19T13:30:51Z | |
dc.date.available | 2014-12-19T13:30:51Z | |
dc.date.created | 2010-09-02 | nb_NO |
dc.date.issued | 2005 | nb_NO |
dc.identifier | 346749 | nb_NO |
dc.identifier | ntnudaim:1079 | nb_NO |
dc.identifier.uri | http://hdl.handle.net/11250/250125 | |
dc.description.abstract | In this paper we look at ways to include context into link analysis algorithms used by search engines to help rank results. We review two basic link analysis algorithms, and propose two methods to include context into the algorithms, combining to a total of 6 different algorithms. We have implemented the algorithms in a domain, and conducted experiments with a number of users to see which algorithms produce the best results. Our conclusion is that context can help improve the results of the link analysis algorithms, but one has to take great care when implementing it, as there are many pitfalls of using context. | nb_NO |
dc.language | eng | nb_NO |
dc.publisher | Institutt for datateknikk og informasjonsvitenskap | nb_NO |
dc.subject | ntnudaim | no_NO |
dc.subject | SIF2 datateknikk | no_NO |
dc.subject | Program- og informasjonssystemer | no_NO |
dc.title | Experiments on Context-Aware Link Analysis Algorithms | nb_NO |
dc.type | Master thesis | nb_NO |
dc.source.pagenumber | 82 | nb_NO |
dc.contributor.department | Norges teknisk-naturvitenskapelige universitet, Fakultet for informasjonsteknologi, matematikk og elektroteknikk, Institutt for datateknikk og informasjonsvitenskap | nb_NO |