Abstract
Advancement in the field of communication facilitates the world for electronic sharing of huge amount of data. Relational databases most widely used and shared through Internet by the communities and organizations. They work in collaborative environment for sharing of knowledge and for applying data mining techniques for finding different market trends and interests for research groups. Proof of ownership and database security are major issues related to online sharing of database. Watermarking facilitates solution for ownership protection and reversible watermarking ensures data recovery in case of any kind of malicious cyberattacks. The major challenge of watermarking are (1) imperceptibility of watermark, (2) robustness of watermark against any kind of data alteration attacks, and (3) data recovery. In this research, a reversible watermarking technique for numerical relational databases by using evolutionary techniques has been proposed that ensures the integrity of underlying data and robustness of watermark. Moreover, maximum Relevance and Minimum Redundancy (mRMR) technique has been used for feature subset selection. Binary Bat Algorithm (BBA) as constraints optimization technique used for watermark creation. Result shows effectiveness of the proposed scheme against data tempering attacks. The proposed BBA provides better results as compare to Genetic Algorithm.