Sigma Synthetic

Sigma Synthetic analyzes every dimension of a consumer identity to detect whether the identity is real or synthetic. The product consists of the Sigma Synthetic model, which provides a score based on the likelihood of a given identity being synthetic, and a list of reason codes that provide insight into risky and mitigating factors. Powered by a machine learning first methodology, the model uses a rich set of feedback that includes both identity and behavioral data. This allows the solution to be used across the customer lifecycle to verify an identity.

The Sigma Synthetic product targets the detection of both manipulated identities (often referred to a first-party synthetic fraud) and fabricated identities (often referred to a third-party synthetic fraud).

How it works

The Sigma Synthetic product returns the following outputs:

Synthetic Risk Scores

The synthetic risk score is a probabilistic value between 0.001 and 0.999 that predicts the inherent risk of the input identity being synthetic. A higher score indicates a greater likelihood of the identity being synthetic.

The scores are uniformly normalized across Socure's network. This means that, a score of 0.800 indicates that the consumer is more likely to be synthetic than 80 percent of all consumers, but less likely to be synthetic than 20 percent of consumers. In general, consumers with scores equal to or greater than 0.990 are considered highly likely to be synthetic.

For most integrations a single score will be returned in the response, but more than one score can be returned. This will be coordinated with your account team when it is necessary, and is generally used during upgrades to newer model versions.

📘

Note:

In addition to the model score, a name and a version will also be returned for each score. These values are returned for auditing and to differentiate between models when multiple scores are returned.

Reason Codes

The reason codes in the response provide detailed information to help you understand the factors that influenced the synthetic risk score. Reason codes are boolean indicators of conditions that were applicable to the consumer and can be used to refine decisioning workflows.

Two important reason codes are R298 and R299 which indicate whether a synthetic identity appears to be a manipulated or a fabricated identity. This information should be used to inform step-up actions:

• R298: Identity resembles a manipulated Synthetic identity - step up to eCBSV if this reason code is returned together with a high Synthetic score.
• R299: Identity resembles a fabricated Synthetic identity - step up to [Document Verification](doc:Doc V) if this reason code is returned together with a high Synthetic score.

See Reason Code in the RiskOS™ Dashboard for information on which additional reason codes can be returned as part of the Sigma Synthetic product.