How is "Predicted to be Popular" calculated in Rialto
Predicted to be popular, Rialto, Ranking profiles, Ranking boosts
Question
How is Predicted to be Popular calculated in Rialto?
Answer
Rialto uses measures of historical sales of the title and two data models of purchase patterns for books sharing similar document metadata.
Rialto compares fourteen fields, out of which eight are derived from LCC and Dewey.
We use a combination of the overall predicted sales value and representation of metadata preferences at customer level, to predict customer-specific sales value. To achieve this, Rialto uses the above fourteen fields, while leveraging mutual information on holdings and usage.
For more information on Ranking Profiles and Ranking Boosts see Managing Ranking Profiles.
If you have any questions, please contact us via live chat or submit a support ticket here.