Recommendations have become an expected part of the customer experience. Internet leaders such as Amazon and Netflix have set this bar very high. Quality recommendations help individuals find things they may not have thought of themselves. This positive buyer experience has become a critical tool for online retailers, the hospitality industry and countless other service-driven organizations.
Hadoop provides a scalable, flexible and cost-effective platform needed to build a Big Data Recommendation engine. Leveraging the collaborative filtering and clustering libraries within Mahout, development of a basic yet powerful recommender is very straightforward.
In this class, we will discuss some of the key algorithms used in common recommenders. We will then walk through the critical design aspects of building a recommendation engine and share tips and leanings from real-world implementation. We will then walk through the code and operation of a functioning Big Data recommender, available as a fully functional virtual machine distributed during the lecture.
Note: We will interactively demonstrate a few key components of the sample application. Although not a requirement, we encourage attendees to download our Hadoop VM and clone the latest version of our sample project. The Hadoop VM is over 1GB, and a stable release, please download prior to the class.
Download the Hadoop Virtual Machine
and Sample Recommender Github repository