4. Case study of a recommendation platform: Mahout

Mahout is a platform supported and developed by the Apache community. The aim of the Mahout project is to provide an environment for rapidly creating robust, high-performance learning algorithms. The platform integrates a large number of algorithms, including recommendation algorithms. To improve processing and facilitate scaling, the algorithms can take advantage of the Map/Reduce paradigm, for example.

The architecture of a recommendation application based on Mahout is built on four main layers (figure 11 ):

• the application that will use the recommendation process;

• the recommendation method...