Presentation of Offered Services: Babelio, A Recommendation Engine Dedicated to Books
Vassil Stefanov, Guillaume Teisseire, Pierre Frémaux · 2014
With the emergence of digital mediation in general and the e-commerce of e-books in particular, new recommendation mechanisms are necessary and help to point the consumer of cultural products toward the potential “premium product”. In the case of Babelio, its recommender systems are based on an algorithmic and social double approach, which is described in this chapter. An automated recommendation engine requires an organization into a hierarchy by qualitative pertinence of the list of recommended titles. One of the issues of recommendation engines is favoring the discovery of titles, and designing this according to the preferences of the reader. The performance issue is linked to the need to build stable systems with a controlled consumption of resources. The issue of scalability, more generally, characterizes the uncertainty of a recommendation method used in a network of 10,000 readers remaining pertinent in a network of a million readers.