Recommender Systems - One Day Short Course
How do companies like Amazon, Netflix and eBay decide what products and links to suggest? Recommender systems are algorithms that incorporate past user behavior, query strings, product descriptions and product reviews to produce highly targeted recommendations. In this short course, you'll learn about the techniques and algorithms available for these sorts of problems and how your problem statement might affect the choice of algorithms.
This class is aimed at computer scientists interested to learn about recommender systems. No previous experience with machine learning is required. You'll see code examples for the various algorithms and then use this code to build your own recommender systems in a variety of different problems. We'll use R programming language for running algorithms and will provide a brief introduction to R in order to get everyone up to speed.
1. Introduction to R programming Language
2. Background on Recommender Systems
3. Content based recommendations
4. Transactions based recommendations
Association Rule approach
5. Scaling up
Payment and Refund
The class will cost $210 with early registration. For early registration go to http://recommendersystems.eventbrite.com/ or you can use paypal (mike at mbowles dot com). You can register and pay by check or cash the day of the event. Eventbrite payments are refundable up to two days before the event. We will have a strict 20 person limit on live attendance. Meetup signup doesn't guarantee a seat.
In class, we'll run through numerous code examples together. I'll pass out the code at the beginning of class. For you to execute it with me, you'll need to install R and RStudio. Here are links where you can download installers for your OS.
In class, we'll add the recommenderlab package to the base R installation.
We'll take a lunch break, so bring a bag lunch with you.