Enhancement of k-anonymity algorithm for privacy preservation in social media

Abstract

In recent times, more and more social data is transmitted in different ways. Protecting the privacy of social network data has turn out to be an essential issue. Hypothetically, it is assumed that the attacker utilizes the similar information used by the genuine user. With the knowledge obtained from the users of social networks, attackers can easily attack the privacy of several victims. Thus, assuming the attacks or noise node with the similar environment information does not resemble the personalized privacy necessities, meanwhile, it loses the possibility to attain better utility by taking benefit of differences of users’ privacy necessities. The traditional research on privacy-protected data publishing can only deal with relational data and even cannot applied to the data of social networking. In this research work, K-anonymity is used for providing the security of the sensitive information from the attacker in the social network. K-anonymity provides security from attacker by making the graph and developing nodes degree. The clusters are made by grouping the similar degree into one group and the process is repeated until the noisy node is identified. For measuring the efficiency the parameters named as Average Path Length (APL) and information loss are measured. A reduction of 0.43% of the information loss is obtained.

K-Anonymity Enhancement for Privacy Preservation with Hybridization of Cuckoo Search and Neural Network using Clustering

“…In the initial segment, the outcome of the proposed work has been delineated and the subsequent section has shown the comparison of proposed work with the conventional techniques. For the comparison, the research work of [19] and [23] has been considered that has utilized APL and Information loss as the computing measures.…”

“…Social networking privacy preservation without disclosing the sensitive information of users is an imperative issue. Clustering based technique clusters the edges and the vertices in groups and then anonymised the subgraph in super-vertex [19]. Accordingly, the details of the individuals could be hidden appropriately.…”

“…Aanchal Sharma and Sudhir Pathak (2018) [19] have utilized k-anonymity for the privacy preservation of sensitive information in social networking. The clustering concept has been used by combining the same degree in one group and the procedure is iterated till there is an identification of noisy node.…”

“…This section elaborates the comparison of proposed mechanism with the existing mechanism to depict the effectiveness of the work. For the comparison, [19] and [23] has been considered. Table 7.Comparison of APL for proposed and existing work [19] Table 7.Comparison of APL for proposed and existing work [19] Table 7.Comparison of APL for proposed and existing work [19] Existing [19] Table 7.Comparison of APL for proposed and existing work [19] 0.15 Table 7.Comparison of APL for proposed and existing work [19] 0.22 Table 7.Comparison of APL for proposed and existing work [19] 0.36 Table 7.Comparison of APL for proposed and existing work [19] 0.35 Table 7.Comparison of APL for proposed and existing work [19] 0.37 Table 7.Comparison of APL for proposed and existing work [19] 0.36 Table 7.Comparison of APL for proposed and existing work [19] 0.34 Table 7.Comparison of APL for proposed and existing work [19] 0.31 Table 7.Comparison of APL for proposed and existing work [19] 0.28 Table 7.Comparison of APL for proposed and existing work [19] 0.41 More information loss leads to less reliability of the novel mechanism, so, it should be less for more privacy of the system.…”

“…The comparison of APL for proposed and existing work [19] for k=10 has been delineated in Fig.8 and is shown in tabular form in table 7. There is a reduction of 14.6% in APL of the proposed work as contrasted to existing work.…”