Using graph clustering, we improve accuracy of Kronecker social networks which are protected by differential privacy. Ensuring the differential privacy implicates addition of marginal changes to the ...