Anonymous opt-out and secure computation in data mining

by Shepard, Samuel Steven.

Abstract (Summary)
ii Ray Kresman, Advisor Privacy preserving data mining seeks to allow users to share data while ensuring individual and corporate privacy concerns are addressed. Recently algorithms have been introduced to maintain privacy even when all but two parties collude. However, exogenous information and unwanted statistical disclosure can weaken collusion requirements and allow for approximation of sensitive information. Our work builds upon previous algorithms, putting cycle-partitioned secure sum into the mathematical framework of edge-disjoint Hamiltonian cycles and providing an additional anonymous “opt-out” procedure to help prevent unwanted statistical disclosure. iii
Bibliographical Information:


School:Bowling Green State University

School Location:USA - Ohio

Source Type:Master's Thesis

Keywords:data mining computer networks hamiltonian systems


Date of Publication:

© 2009 All Rights Reserved.