Title: A novel framework for optimised privacy preserving data mining using the innovative desultory technique
Authors: J. Indumathi, G.V. Uma
Addresses: Department of Computer Science and Engineering, Anna University, Chennai 600025, Tamilnadu, India. ' Department of Computer Science and Engineering, Anna University, Chennai 600025, Tamilnadu, India
Abstract: Computing is an ogle spectator witnessing afloat similar to a tornado. The chic data analysis and mining techniques compromise privacy and their exploitation have reached the pinnacle, demanding our attention. Amongst the available diverse techniques available for privacy preservation, the existing techniques involve only individual preservations. This work proposes novel three-tier architecture. The subsequent coalescing of three best secure techniques endows us with a three-fold privacy preservation, namely access control limitation technique, randomisation and Privacy Preserving Clustering (PPC). Thus, our framework gives an efficient control system comprising authentication, authorisation and access for each database application; efficiency and economical benefits of randomisation and advantages of PPC. We have further proposed a method for computing the object-based dissimilarity for secure computation for all different attribute types. PPC uses a combination of CLARANS with Simulating Annealing in order to achieve increased privacy, efficient scaling and improvised performance.
Keywords: access control; CLARANS; clustering; data mining; desultory; dissimilarity matrix; PPDM; privacy preservation; randomisation; simulating annealing; privacy protection; control limitation; privacy preserving clustering; authentication; authorisation; computer security.
International Journal of Computer Applications in Technology, 2009 Vol.35 No.2/3/4, pp.194 - 203
Published online: 20 Jun 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article