10-04-2017, 09:04 PM
m-protection for Collaborative publication Data
Abstract:
"We believe the joint decision of publishing data to anonymous data to multiple sections horizontally data providers. Let's look at a new type of "insider attack" by the collusion of data providers, who may use their own data recording (a subset of the total data), in addition kvne nij background knowledge to write data from other data providers. the paper examines this new threat and makes some contributions. First, we introduce the notion of m-privacy, which ensures that anonymous data meet confidentiality restrictions specified for any group to m conspiracy data providers. Second, we present a heuristic algorithms use of group equivalence of the monotony of the Vasi constraints and adaptive methods to streamline for the effective control of m-privacy is given a set of records. Finally, we present the data provider known algorithm of anonymization with Adaptive m-privacy strategies for ensuring a high usefulness and m-privacy anonymous data with efficiency. Experiments on real data sets show that our approach provides better or comparable utility and effectiveness over existing and original algorithms, providing guarantee of m-privacy. "