Abstract: Efficient and reliable data hiding protection can effectively prevent the active data disclosure of power network data center. Current data hiding methods suffer from high redundancy and low sparse. A data hiding protection method for power grid based on content sensation cluster is proposed, which constructs support vector machine data extract features framework. Orthogonal regression data clustering parameter are designed to control the data hiding number. Simulations and tests demonstrate the improved performance of new method.

Key words: data hiding, data clustering, data center, sparse