Attribution Confidence

Attribution confidence is a measure of how reliable a given entity attribution or cluster assignment is, based on the strength of the evidence behind it.

Attribution confidence acknowledges that blockchain analytics is inferential, not certain. A cluster built from common-input ownership carries higher confidence than one leaning on change-address detection or wallet fingerprinting; a label from a directly observed deposit outranks one inferred from behavior. Expressing this as a confidence value lets downstream systems weight evidence rather than treat every attribution as fact. A compliance rule might act on a high-confidence sanctioned label but merely flag a low-confidence one for review. Because exposure breakdown and address risk scores inherit the confidence of the labels beneath them, propagating it honestly prevents overstated certainty.