We constantly publish updates to our documentation, which may not yet be available in your language. For the most current information, use the English version.

for Partners

Dynamic network integration•4. Set up S2S measurement•S2S impressions with probabilistic modeling

Record S2S impressions with probabilistic modeling

Probabilistic modeling is Adjust's secondary attribution method, and uses machine learning to support a statistical approach to measurement.

When you send certain information to Adjust with an S2S request, Adjust can use this to attribute impressions using probabilistic modeling. Below is a table of the parameters you should send to Adjust.

If you have any questions around the user_agent and how to test it, please reach out to integration@adjust.com.

Important:

If the user_agent is missing from an S2S impression, Adjust does not use probabilistic modeling for attribution. However, the impression is still counted in Datascape.

Parameter	Example value
`user_agent`	Mozilla%2F5.0+%28Linux%3B+U%3B+Android+4.4.2...
`ip_address`	8.8.8.8