Reduced-parameter approaches to time-sensitive discovery and analysis of navigational patterns

Simple item page
dc.contributor.advisorBarker, Kenneth E.
dc.contributor.authorUdechukwu, Ajumobi Okwuchukwu
dc.date.accessioned2017-12-18T21:08:15Z
dc.date.available2017-12-18T21:08:15Z
dc.date.issued2006
dc.descriptionBibliography: p. 130-142en
dc.description.abstractNavigational pattern discovery is the application of pattern mmmg techniques to navigational studies. Previous work on navigational pattern discovery strongly relies on user-specified preset parameters (notably, support thresholds). A major drawback of parameter-driven data mining is setting appropriate thresholds. In most practical scenanos, data mining 1s iterative, meaning that the analyst may mme at varymg thresholds before arriving at satisfactory results. This usually proves expensive m parameter-driven techniques. These challenges faced in parameter-driven data mining become very pronounced in environments with continuously streaming data where it is no longer possible to apply simple iterative data mining techniques. This thesis explores a new design goal for the discovery and analysis of navigational patterns, which aims to reduce the use of pre-defined parameters required at the beginning of the mining process (reduced-parameter mining), and completely removes pre-defined parameters wherever possible (parameter-free mining). The new paradigm is developed in the context of time-sensitive navigational pattern discovery. Three broad dimensions are explored where time can significantly affect the ways navigational patterns are to be discovered or analyzed. The first dimension 1s m information systems where the content continuously changes with time (thus affecting navigational behaviour). The second dimension is in high-volume information systems, where the usage-logs continuously change with time. The final dimension of temporal significance in navigational pattern discovery explored in the thesis is the representation and analysis of navigational patterns as full temporal objects ( or time series). This thesis conceptualizes the notions of time significance in navigational pattern discovery with respect to the dimensions identified above, and then proposes novel techniques for discovering navigational patterns in such environments that are based on reduced-parameter or parameter-free principles. Interestingly, the results from this thesis show that reduced-parameter mining and parameter-free mining are practical concepts. The results also show that reduced-parameter/parameter-free mining techniques out­perform parameter-based mining techniques in environments that require continuous updates of patterns. This is contrary to the general belief that early introduction of pre­defined parameters always improves the mining process. This result has strong implications to other data mining tasks.
dc.format.extentxiii, 142 leaves : ill. ; 30 cm.en
dc.identifier.citationUdechukwu, A. O. (2006). Reduced-parameter approaches to time-sensitive discovery and analysis of navigational patterns (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/437en_US
dc.identifier.doihttp://dx.doi.org/10.11575/PRISM/437
dc.identifier.urihttp://hdl.handle.net/1880/101438
dc.language.isoeng
dc.publisher.institutionUniversity of Calgaryen
dc.publisher.placeCalgaryen
dc.rightsUniversity of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission.
dc.titleReduced-parameter approaches to time-sensitive discovery and analysis of navigational patterns
dc.typedoctoral thesis
thesis.degree.disciplineComputer Science
thesis.degree.grantorUniversity of Calgary
thesis.degree.nameDoctor of Philosophy (PhD)
ucalgary.item.requestcopytrue
ucalgary.thesis.accessionTheses Collection 58.002:Box 1687 520492204
ucalgary.thesis.notesUARCen
ucalgary.thesis.uarcreleaseyen
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
thesis_Udechukwu_2006.pdf
Size:
63.18 MB
Format:
Adobe Portable Document Format
Description:
Thesis
Download
Collections
Legacy Theses