Practical Machine Learning: Innovations in Recommendation

by Ted Dunning, Ellen Friedman

Building a simple but powerful recommendation system is much easier than you think. Approachable for all levels of expertise, this report explains innovations that make machine learning practical for business production settings—and demonstrates how even a small-scale development team can design an effective large-scale recommendation system.

Apache Mahout committers Ted Dunning and Ellen Friedman walk you through a design that relies on careful simplification. You’ll learn how to collect the right data, analyze it with an algorithm from the Mahout library, and then easily deploy the recommender using search technology, such as Apache Solr or Elasticsearch. Powerful and effective, this efficient combination does learning offline and delivers rapid response recommendations in real time.

Understand the tradeoffs between simple and complex recommendersCollect user data that tracks user actions—rather than their ratingsPredict what a user wants based on behavior by others, using Mahoutfor co-occurrence analysisUse search technology to offer recommendations in real time, complete with item metadataWatch the recommender in action with a music service exampleImprove your recommender with dithering, multimodal recommendation, and other techniques

Genres: Nonfiction, Technology, Ebooks, Computers, Technical

59 pages, Kindle Edition

First published April 16, 2014

Reviews

[Luke]

This is a very high level overview - nothing in depth. You can skim read this in less than a couple of hours. If you have next to no technical knowledge and have barely heard of a recommender system, then this is the book for you. Otherwise, I'd suggest saving your time by not reading this.

[Siarhei]

This short book feels like an excerpt of a one-day technical course, “Machine Learning with Apache Mahout: Introduction to Scalable ML for Developers” that they promote on the last pages.

[Mike Fowler]

Interesting concepts but not much depth in explanations.

[Dr. Tathagat Varma]

Discusses some foundational ideas in reocmmendation.

[James Tomasino]

All but useless high level info.

[ZeV]

A very high-level overview of how a recommendation system could be built with a Hadoop-based batch processing with a Lucene-based search API. I think the dependence on Hadoop, which simply isn't easy to manage for most development teams, makes this an obsolete recommendation.

[Daron Yondem]

This was indeed practical. It's a very short read providing a lot of insights and tricks in ML for a newbie.

[Randy]

Really short book. Sort of helpful to read a mention of a concept one more time, and the engine idea is ok.

I'd only recommend the book because it's so short it wouldn't take up much time.