Probabilistic Modeling | AppsFlyer

Probabilistic Modeling

AppsFlyer’s probabilistic modeling method uses advanced statistical algorithms, based on machine learning, to project the likelihood that an app install was generated by an engagement with a particular campaign. This method is utilized for landscapes where deterministic attribution isn’t an option.

Probabilistic modeling doesn’t generate any ID for device or user; it’s a completely anonymized method for statistical inference.

« Back to Glossary Index