Clustering-based approach for private Chain data anonymization

Abstract

Blockchain is a new type of distributed data storage technology. With the rapid development of blockchain, vast amounts of data have been accumulated in these applications, which provide researchers with unprecedented opportunities to analyze blockchain data. However, if blockchain data is published openly, it may cause privacy leaks. Owing to the characteristics of blockchain data, the traditional anonymous method based on data publishing cannot be directly applied to blockchain data. This paper proposes an anonymous method based on clustering named clustering partition based on Bisecting k-medoids (CP-BK). We treat the transaction data of the blockchain as table data and use the k-anonymity model to protect the identity privacy of blockchain users. Finally, we evaluated the information loss and efficiency of this algorithm in experiments.