Tamper Detection of Relational Database Based on SVR Predictive Difference

Database authentication verifies valuable content from illicit copying, illegal redistribution and maliciously tamper while it is to be sold or be outsourced to un-trusted parties. This paper exploits the digital watermarking technology for guaranteeing the database integrity underlying distortion free of database content. The proposed scheme employs SVR predictive function to obtain characteristic of the database and uses Huffman coding to encode the characteristic for compressing important payload information. In detection phase, minor and necessary additional payload information of the database is used to accomplish tampering detection.