In this chapter, we saw how to build recommendations using Apache Mahout. We looked at how we can leverage Mahout for both the standalone and the distributed mode. We have written Java code for user-based, item-based, and SVD-based recommendation engines in the standalone mode and Alternating Least Squares recommendations in the distributed mode. We also saw how we can evaluate the recommendation engine models. In the final section, we explored a very basic system of how to take Mahout to production.
In the final chapter, we shall cover the future of recommendation engines, where the recommendation engines are heading, and the promising use cases to lookout for.