Traceable high-dimensional data publishing based on Alliance Chain

High-dimensional data publishing based on Local Differential Privacy (LDP) mainly adopts the method of synthesizing similar datasets, which only needs to rely on semi-honest centers. LDP disturbs the data locally to avoid the illegal leakage of data by the data collector. However, after the data is published, the data receiver may illegally transmit the private dataset, and the existing work cannot realize the responsibility tracing and copyright protection. Besides, the existing work cannot solve the problem of false datasets published by data collectors. In this paper, the reversible robust watermark is used to embed the identity information of the data receiver into the generated database. Once the leakage occurs, the watermark can be extracted to trace the source, which solves the problem of illegal transmitted by the data receiver. In addition, the consensus mechanism of Alliance Chain was used to make the dataset published by the data collector be approved by most members of the chain, which solved the problem of false release by data collector. Experiments show that our method has good usability and scalability. In terms of robustness, even if the attacker tampers more than 80% of the data, it can still ensure the rationality of the dataset.