Social network anonymization is a powerful privacy-preserving technique for large social networks. In this paper we address the problem of community preservation in anonymized social networks. We investigate how well communities are preserved when a specific anonymization model, namely k-anonymity for social network, is used. We use a well-known community detection algorithm based on modularity optimization for our study. To assess the results, we employ three measures (naieve community preservation, community preservation at node level, and normalized mutual information) to quantify community preservation between original and final communities. Our results show that, in general, the community structure is partially preserved in the anonymized networks.
展开▼
