Privacy Protection Method on Publishing Dynamic Set-Valued Data
Jian Zhang, Yu Chieh Yang · 2016
After researching many protection methods for releasing sensitive information, we found that differential privacy could provide strong guarantees on it, but it will adding too much noise and spending too much times in releasing dynamic set-valued data.To solve these problems, this paper present a privacy protection method based on Diffpart.For a dataset need to be published, set it by Diffpart algorithm firstly, then using sampling method to sample some nodes and add the Laplace noise to them to protect the sensitive information when dataset needs to be updated.Then, generate a transposing number randomly to adjust the sampling nodes for subsequent update.Through the experiment, compared with the Diffpart algorithm, the method that we raised reached a ideal effect in practicability and protective for data release.