Clustering User Behavior in Scientific Collections

• Main Author: Blixhavn, Øystein Hoel
• Format: Others
• Language: English
• Published: Norges teknisk-naturvitenskapelige universitet, Institutt for datateknikk og informasjonsvitenskap 2014
• Subjects: ntnudaim:12121; MTDT Datateknologi; Data- og informasjonsforvaltning
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:no:ntnu:diva-27340

This master thesis looks at how clustering techniques can be appliedto a collection of scientific documents. Approximately one year of serverlogs from the CERN Document Server (CDS) are analyzed and preprocessed.Based on the findings of this analysis, and a review of thecurrent state of the art, three different clustering methods are selectedfor further work: Simple k-Means, Hierarchical Agglomerative Clustering(HAC) and Graph Partitioning. In addition, a custom, agglomerativeclustering algorithm is made in an attempt to tackle some of the problemsencountered during the experiments with k-Means and HAC. The resultsfrom k-Means and HAC are poor, but the graph partitioning methodyields some promising results.The main conclusion of this thesis is that the inherent clusters withinthe user-record relationship of a scientific collection are nebulous, butexisting. Furthermore, the most common clustering algorithms are notsuitable for this type of clustering.