Since the dawn of time - or at least since man started using ID+ for consumer verification of individual identities, the AuthScore has been the central component of an ID+ Score. A simple ordinal whole number between 0 and 10, it is the cornerstone of ascertaining if a consumer being verified is the person they claim to be. For the latest relase of Socure ID+ for Consumer Verification, version 2.5, the AuthScore is "new and imrpoved" for better validation.
AuthScore is a component of the data provided from an ID+ Identity Check. Taken together with the fraud score and associated reason codes, AuthScore forms the basis of identity verification used in organizations to validate identities in over 180 countries worldwide.
We’re pleased to announce a significant update to AuthScore available immediately in the July 2016 release of the Socure ID+ v2.5 API! The “new and improved” AuthScore is a much stronger predictor of whether a provided identity is that of a real person versus one that is synthetic or fake.
In order to do this, we are replacing the current AuthScore that is a weighted average of a number of input social and non-social data points with the new AuthScore which is calculated by using a supervised model. The combination of mathematical models (supervised, unsupervised and computational) derives the AuthScore by weighting data from the following three areas:
- PII correlation: Derived intelligence on how well the components of the provided identity are correlated - see more about this in our recent post on correlation scores.
- Digital presence: Checks for the presence of profiles on a large number of social and non-social sites online to ascertain if the components of the provided identity match known social profiles.
- Online behavior: Verifies how closely the behavior of the profile identified by the provided identity resembles the behavior of a real person with an established history on such sites.
AuthScore is also an input into the FraudScore model, which additionally looks at a host of other predictive risk indicators of fraud. The current FraudScore model has be trained from known fraud in the millions of identities verified with Socure and the Social Biometrics platform. This data on “known bad” entities is fed back into our advanced machine learning algorithms, to enable The Social Biometrics™ Platform to get increasingly smarter about how to predict whether an identity is likely being used in a fraudulent manner.
AuthScore had broad utility as a generic indication of authenticity, and has proven useful in the around the world across numerous industry segments and applications.
To learn more about how Socure works, contact your representative or firstname.lastname@example.org, you can also click the button below to set a meeting with us: