Getting Started with Recommendation Systems

Next Monday’s Data Science DC Meetup is on Recommendation Systems in the Real World, and features two outstanding local data scientists, Matt Bryan from WaPo Labs and Bryce Nyeggen from LivingSocial. To preview that event, we asked them to suggest some resources for people interested in learning about this important technology. Recommendation engines are the systems that underlie personalization features such as Netflix’s movie suggestions and Amazon’s product suggestions.

Both Matt and Bryce suggested Mahout in Action as a good, practical book on the topic. Apache Mahout is a Java-based system for building scalable machine learning applications, and it supports several technologies used in recommendation engines.

For those not interested in Mahout, Programming Collective Intelligence uses Python instead of Java, and is a good introduction to recommendation systems, despite being 5 years old at this point. Machine Learning for Hackers is another option, with a short chapter on a simple recommendation algorithm written in R.

Once you’ve gotten started, you’ll want to start experimenting with some standard data sets. See last week’s Data Source Weekly for some recommendations, then come to Monday’s Meetup for some real-world wisdom!

The following two tabs change content below.
Harlan Harris has a PhD in Computer Science (Machine Learning) from the University of Illinois at Urbana-Champaign, and post-doctoral work in Cognitive Psychology at several universities. He currently is Senior Solutions Architect at Sentrana, Inc., and co-organizes Data Science DC.

Latest posts by Harlan Harris (see all)

This entry was posted in Data Science DC. Bookmark the permalink.

One Pingback/Trackback