Clustering-Based Anonymization Technique using Agglomerative Hierarchical Clustering

Abstract

Privacy protection has emerged as one of the most attractive research topics in recent years. Today; with the rapid growth of social networks (SNs), the rate of users’ data collection has increased. Users’ sensitive information of collected data should be protected from adversaries. One of the most commonly used privacy protection techniques is data anonymization. Data anonymization is achieved by changing or deleting some information. By applying all three commonly accepted constraints; K-anonymity, L-diversity, and T-closeness; we introduced a data anonymization strategy based on agglomerative hierarchical clustering to preserve anonymized data from identity disclosure, attribute disclosure, and similarity attacks. The simulation result over the dataset of two popular SNs illustrates the effectiveness of the proposed algorithm.