دورية
Privacy preserving using joint 2 K-means clustering and coati optimization algorithm for online social networks
| العنوان: | Privacy preserving using joint 2 K-means clustering and coati optimization algorithm for online social networks |
|---|---|
| المؤلفون: | Gowda, Nithish Ranjan, Venkatesh, Venugopal, K. R. |
| المصدر: | International Journal of Information Technology; April 2024, Vol. 16 Issue: 4 p2715-2724, 10p |
| مستخلص: | Social networks, that have grown so prevalent nowadays, enable users to exchange information and save a significant amount of personal data about them. Although the information that is stored can be useful for enhancing the quality of services provided to users, it also poses a risk to their privacy. This is because social networks contain private information about users. As a result, members of social networks seek to protect the confidentiality of their shared data. The most popular method for protecting confidentiality is anonymizing data, which involves modifying or eliminating some information while trying to preserve as much of the original data as possible. The high level of data loss, similarities attacks, and protection from attribute or link disclosure is problem with existing anonymity approaches. The study proposes a hybrid approach based on K-member k-means clustering and coati optimization algorithm (2KMCOA) as a successful solution for balanced clustering and anonymizing in social networks in order to get over these shortcomings. A K-member K-means clustering algorithm is used to divide the different users into C clusters and each cluster has at least K users, as part of the proposed anonymization procedure. Following clustering, an initial solution is generated that produces the modified data table which should satisfy objective functions and three constraints. The coati optimization algorithm (COA) is then utilized to optimize the primary clusters even more to anonymize the data as well as network graph. The efficiency of the proposed 2KMCOA is compared with other existing anonymity techniques K-means clustering with COA (KMCOA) and K- member K-means clustering without COA (2K) in terms of clustering error, balancing error, distortion rate, objective function, cost function and CPU running time. |
| قاعدة البيانات: | Supplemental Index |
| تدمد: | 25112104 25112112 |
|---|---|
| DOI: | 10.1007/s41870-024-01729-w |