Address Clustering
Grouping blockchain addresses that appear to be controlled by the same person or company so activity can be counted per entity.
Clustering rests on heuristics. The common-input rule assumes that addresses whose funds are spent together in one transaction share an owner, which holds on UTXO chains except where users deliberately combine funds with strangers. Change-output detection identifies which output returns to the sender. Deposit-address linking connects customer addresses to an exchange by following where funds are consolidated. Public disclosures, court filings, and investigations supply named labels for the largest wallets. Every entity-level metric, from exchange balances to holder concentration, depends on this layer, and it is probabilistic: coin-mixing tools, batched exchange payouts, and shared custody produce false links, and clusters are merged or split later, which quietly rewrites history.
Dalam praktik
Exchange balance charts exist only because analytics firms maintain labeled address sets for each exchange, which the exchanges themselves usually do not confirm.
Kesalahpahaman yang umum terjadi
Clustering is inference rather than fact: entity-level metrics rest on heuristics that can be defeated deliberately and are revised when a label proves wrong.