Nearest neighbor tour circuit encryption algorithm based random laplacian eigenmap

Abstract

This paper presents nearest neighbor tour circuit encryption algorithm based random Laplacian Eigenmap. In order to be suited for privacy-preserving classification, we first alter the selection fashion of the parameters nearest neighbor number k, embedded space dimension d and heat kernel factor t of Laplacian Eigenmap algorithm. Further we embed the tourists' sensitive attribution into random dimension (even higher) space using random Laplacian Eigenmap, thus the sensitive attributes are encrypted and protected. Because the transformed space dimension d and the nearest neighbor number k are both random, this algorithm is not easily be breached. In addition, Laplacian Eigenmap can keep topology structure of dataset, so the precision of classification after encryption are not affected. The experiment show that the present method can provide tourists' sensitive information enough protect, and it also give the tourist appropriate tour circuit. © 2009 IEEE.