ROTATION BASED TRANSFORMATION METHOD AND APPARATUS FOR PRESERVING DATA PRIVACY

摘要

The present invention comprises a rotation-based transformation methods for protecting the privacy of the data. Rotation-based conversion method of the present invention for a given data set is divided into a plurality of data subsets, to apply a different rotation matrix for each of the divided data sub-set to perform the data rotation, the rotation is again junction data subset to obtain a rotated data set. The thus obtained data set is rotated is used to release the data clustering. Such rotational-based conversion method of the present invention is particularly useful in alleviating the AK-ICA attack. On the other hand, when the two parties according to the present invention, parameters such as the number of the sub-data is set to be agreed between the two parties. Data subsets are preferably divided into equal size. In addition, the rotation matrix between the different data subsets using rotation to provide the effectiveness of the clustering, the rotation to remove a plurality of different relative rotational deviation between the rotation matrix unity operations can be performed.