This thesis contains a critical evaluation of the unrealization approach to privacy preserving data mining. We cover a fair bit of ground, making numerous contributions to the existing literature. First, we present a comprehensive and accurate analysis of the challenges posed by data mining to privacy. Second, we put the unrealization approach on firmer ground by providing proofs of previously unproven claims, using the multi-relational algebra. Third, we extend the unrealization approach to the C4.5 algorithm. Fourth, we evaluate the algorithm's space requirements on three representative data sets. Lastly, we analyse the unrealization approach against various issues identified in the first contribution. Our conclusion is that the unrealization approach to privacy preserving data mining is novel, and capable of addressing some of the major challenges posed by data mining to privacy. Unfortunately, its space and time requirements vitiate its applicability on real-world data sets.
Identifer | oai:union.ndltd.org:uvic.ca/oai:dspace.library.uvic.ca:1828/3156 |
Date | 08 December 2010 |
Creators | Williams, James |
Contributors | King, Valerie D., Janhke, Jens H. |
Source Sets | University of Victoria |
Language | English, English |
Detected Language | English |
Type | Thesis |
Rights | Available to the World Wide Web |
Page generated in 0.0016 seconds