Graph-Clustering Anonymity Privacy Protection Algorithm With Fused Distance-Attributes

Abstract Clustering anonymity is a common social network data privacy protection scheme, which based on graph-clustering. Many existing graph clustering methods mainly focus the relationship between structure and attributes of nodes, difference them due to metric usually causes problem poor results. To address shortcomings in above graph-clustering methods, method implemented with fused distance-attributes (GCA-DA) proposed. Firstly, algorithm quantifies distance attribute similarity nodes separately balances differences calculate integrated similarity. Then all are clustered into clusters according two each contains no fewer than k nodes. Finally, anonymized. In this method, generalization for every cluster can prevent attacks by background knowledge attributes. addition, divided numerical non-numerical measure separately, therefore maintain usability better. Experiment results demonstrate effectiveness improving quality reducing information loss.