K-anonymity enhancement for privacy preservation with hybridization of cuckoo search and neural network using clustering

Abstract

Expansion of social network and the publication of its data have directed the risk of disclosure of individuals’ confidential information. Privacy preservation is a must thing before service providers publish the network data. In recent years, privacy in social network data has become the most concerned issue as it has gripped our lives in a dramatic manner. Numerous anonymization methods are there that assists in privacy preservation of social networking and among all, k-anonymity is the utmost one that helps in providing the security by developing graph and nodes degree. In this manuscript, the enhancement of K-anonymity has been addressed with major changes in node editing methodology. The clusters are developed with the integration of the same degree in one group and the procedure is iterated till the identification of noisy data. An advanced Cuckoo Search is commenced for minimizing the node miss placement in groups. The results of the Cuckoo Search are integrated with Feed Forward Back Propagation Neural Networks to cross-check the structure and to reduce the node miss placement in groups. Average Path Length (APL) and Information parameters are measured for the evaluation and comparative analysis and the effectiveness of the research has been checked by comparing the results of Aanchal Sharma and P. R. Bhaladhare. There is a diminution of 14.6% while comparing APL with Aanchal Sharma and 8.61% and 10.38% of reduction is shown with Aanchal Sharma and P. R. Bhaladhare for Information loss.