Anbefalte antall kommer fra Schibsted sin ML modell.
Dokumentasjon vi har fra før om hvordan Aftonbladet henter og poster data fra/til DI: Aftonbladet uses a machine learning model to calculate the circulation of Aftonbladet for each retailer. In collaboration with Aftonbladet, DI has created an integration with their machine learning model. DI shares data the model uses to calculate the circulation, and then the model shares its results through the integration to DI.
An overview of the webservices used by Aftonbladet to send and receive retailer circulation data can be found in this document:
https://docs.google.com/document/d/1vY_YNaIbXcIkFRhIk-vWK13CB4gG5qjHz8G05AJ8JlQ/edit
/wiki/spaces/DI/pages/20939228 - Novus
Returen kommer fra Advantops/Voice Provider løsningen (Tidsam retur
ML endpoints
Logging of novus-actions are logged to LsNovusLog and errors are logged to LsNovusErrorLog. Entries older than 2 days are truncated when the end endpoint is called.
Returns come from the Advantops/Voice Provider solution (Tidsam returns).