Real Time Recommendation Engine

Agenda

Currently deployed recommendation technology almost always provides real-time recommendations based on a model that is developed using off-line techniques. The use of off-line training severely limits the ability to deal with fast changing content such as news or auctions.

I will describe techniques which make it possible to move this training load into true real-time without loss of accuracy. Real-time training of recommendations is rarely done, partly because existing algorithms require periodic off-line training to correct accumulating inaccuracies.

The techniques that I describe do not require off-line training of any kind. In addition, the techniques described are easy to implement using stream processing, noSQL databases and search engines.

Ted Dunning - CHIEF APPLICATION ARCHITECT, MAPR

Ted Dunning is Chief Application Architect at MapR Technologies and committer and PMC member of the Apache Mahout, Apache ZooKeeper, and Apache Drill projects​. Ted has been very active in mentoring new Apache projects and is currently serving as vice president of incubation for the Apache Software Foundation​.​ Ted was the chief architect behind the MusicMatch (now Yahoo Music) and Veoh recommendation systems. He built fraud detection systems for ID Analytics (LifeLock) and he has 24 patents issued to date and a dozen pending. Ted has a PhD in computing science from the University of Sheffield. When he’s not doing data science, he plays guitar and mandolin. He also bought the beer at the first Hadoop user group meeting.

https://www.mapr.com/blog/author/ted-dunning