Monitoring Blockchains with Self-Organizing Maps

Abstract

Blockchains such as those used by the Bitcoin and Ethereum cryptocurrencies provide a global, observable record of all transactions and associated data. Analyzing blockchain data is useful for tasks such as detecting fraudulent activities, studying the use and growth of the system, and understanding its levels of anonymity and traceability. Such analysis is challenging due to the high volume and rapidly changing characteristics of popular blockchains. In particular, online (soft real-time) analysis of blockchains requires methods that adapt organically to changes in the data. This paper describes such a method based on self-organizing maps and reports on experiments using the Bitcoin blockchain data.