Survey on right-protected data publishing with provable Distance-based mining

Abstract: Data exchange and data publishing are becoming an essential part of business and academic practices. Data owners also need to maintain the principal rights over the concern datasets that they share. This survey presents a right-protection mechanism that can provide detectable evidence for the legal ownership of a shared dataset, without compromising its usability under wide range of machine learning, mining, and search operations. It is accomplished by guaranteeing that order relations between object distances remain unaltered. This survey provides mechanisms for establishing the ownership of a dataset consisting of multiple objects. The algorithms also preserve important properties of the dataset, which are important for mining operations, and so guarantee both right protection and utility preservation. In this paper considers a right-protection scheme based on watermarking. Watermarking may distort the original distance graph. The proposed watermarking methodology preserves important distance relationships, such as: the Nearest Neighbors (NN) of each object of the original dataset. It proves fundamental lower and upper bounds on the distance between objects. In particular, it establishes a restricted isometric property, i.e., tight bounds on the expansion of the original distances. This analysis used to design fast algorithms for NN-preserving watermarking that drastically prunes the vast search space.

Keywords: Right-protection, Watermarking methodology, k-NN classification, k-NN preservation.