A zero-watermarking algorithm for relational database copyright protection

To meet the demand of relation database copyright protection, a zero-watermarking algorithm is proposed. we introduce the idea of classification to database watermark working, giving a certain extent ease to the limit of data types and error bounds in database. This algorithm never changes the information of original object, but proves the databases´ copyright belongs through the extraction of data´s characteristics. It´s good at hiding, and without original database in the recover progress, supports blind-checking of watermark information. Experimental result and analysis show the safety and robustness of the algorithm.