Glossary

Vocabulary

These terms are used precisely in our system. Defining them is how an emerging category becomes legible and quotable.

Identity binding

The process of linking a piece of financial evidence to a specific person with a measurable confidence. If the confidence falls below the required threshold, CashKi returns nothing rather than risk attributing evidence to the wrong individual. The threshold itself is a parameter, not a fixed value; it reflects the lender's risk tolerance for misattribution.

Provenance

The complete record of where a piece of financial evidence came from: its source institution, jurisdiction, the period it covers, the method used to verify it, the date it was checked, and how fresh it is. Provenance is not optional metadata — it is the primary output. An attribute without provenance is not an attribute; it is a claim.

Lossy mapping

The condition that arises when a financial concept from one regulatory jurisdiction is translated into a conceptually similar but not equivalent term from another. When a mapping is lossy, the output carries a lossy: true flag and the receiving lender is expected to treat it accordingly. When no defensible equivalent exists, the concept is returned unmapped rather than quietly translated into something that looks familiar but is not.

Predictive validity

The degree to which a financial attribute predicts future credit behaviour in a given population and jurisdiction. For all attributes CashKi currently produces, predictive validity is explicitly stated as unestablished. This reflects the absence of retrospective testing against lender outcomes, not a known absence of predictive power. The field is shown, not hidden. Establishing it requires a retrospective test against real lending decisions.