Year

2008

Degree Name

Doctor of Philosophy

Department

School of Information Systems and Technology - Faculty of Informatics

Abstract

The thesis is that a branch of discrete mathematics, Formal Concept Analysis, when applied to Semantic File Systems can lead to an improved personal information space. Semantic File Systems share many properties with their non semantic brethren, bringing more rich metadata and the ability to directly resolve user queries within the filesystem interface itself.

A filesystem might offer upwards of a million files each of which having in the order of hundreds of discerning attributes. Formal Concept Analysis has typically been applied to a much smaller input data set and there are issues with scalability both in the initial finding of the set of Formal Concepts and also ongoing issues such as finding the list of files which are currently applicable (the extent) for a Formal Concept.

The thesis is largely dependent on improving the scalability of Formal Concept Analysis in order for it to be applied to such a large dynamic data store.

02Whole.pdf (5317 kB)

Share

COinS