Introduction to Recommender Systems

[ghc]

I can't judge the coursera course, but for anyone who is interested in this field and wants a gentle introduction, I high recommend Programming Collective Intelligence (http://shop.oreilly.com/product/9780596529321.do). It covers many of the types of recommender systems that the coursera course is likely to cover, and comes with a lot of nice Python code examples.

It's highly useful knowledge too. I ran across so many startups that needed recommender systems that I launched a company called Algorithmic.ly (http://algorithmic.ly) to help companies without the expertise integrate recommendation systems and other types of algorithms into their projects.

[dougk7]

PCI is a great book. One book I'll recommend after reading it is "Mining of Massive Datasets" (http://infolab.stanford.edu/~ullman/mmds.html) - Chapter 9 especially. It goes into more depth than what's covered in PCI and it's not too difficult to follow.

I'm working on a personal project in recommender systems so I look forward to enjoy this course.

[tungwaiyip]

In this short article I have implemented the algorithm in the Programming Collective Intelligence in 9 lines of code

http://tungwaiyip.info/2012/Collaborative%20Filtering.html

[arms]

I was about to buy PCI a few days ago, but was worried that the information may be out of date, considering it was published six years ago. Do you think most of the material is still pertinent?

[riggins]

Thanks for the recommendation.

[eldog_]

I'd be interested in knowing how much deep learning is changing the algorithms used in this field, given the performance of restricted boltzmann machines on the netflix data set http://www.cs.utoronto.ca/~hinton/absps/netflixICML.pdf.

[elq]

You'd be surprised how overwhelmingly common and effective very simple methods like logistic regression and basic decision trees are for such systems.

Further, RBMs and other deep learning tools require a significantly more sophisticated mathematical background than algebra and a much broader understanding overall.

[Irishsteve]

The netflix prize touched on one of many areas related to recommender systems.

As mentioned already, very simple methods can be really effective. Things such as the UI are also known to have a big impact on how 'useful' people find the recs.

[colincsl]

I am surprised that there is no mention of this in the course syllabus -- in fact it looks like a lot of recent techniques that are missing. They don't even talking about LSA(/SVD)-based methods until the end of the course.

[msellout]

Not much yet. It's not well enough understood yet.

[riggins]

coursera is killing me with courses I want to take.

[DannoHung]

I just wish you could take the classes at your own pace. I'm getting very little out of the "certifications" or whatever. And the TA's/Discussion forums tend towards pretty useless because of how disorganized they are.

[mark_l_watson]

If you want a quick start without taking a class, install the Apache Mahout project - one of the Hadoop map-reduce examples is a recommendation system. You can hack away, and run on Elastic MapReduce if you need to scale. (https://cwiki.apache.org/confluence/display/MAHOUT/Recommend...)

[javindo]

Interesting anecdote: A graduate of my university works for Google who originally had a very complex "machine learning pipeline" for the product recommendations but he has since re-implemented the feature in, as he calls it, a "much simpler bloom filter algorithm".

[victorhooi]

Hmm, seems interesting = I'm currently doing the Machine Learning one also via Coursera, run by Andrew Ng, and it's good gentle introduction to the subject.

It's a shame we can't view the course content for this one earlier...haha.

[pallandt]

I may be wrong, but I think it's the 1st time the course is being offered, I don't personally remember seeing it before in their course list.

[oneeyedpigeon]

I couldn't possibly recommend a site that requires javascript to display any content whatsoever.

[jwr]

Why not? JavaScript has become a part of the web, just as much as HTML and CSS. You might as well shun sites that use the <ul> tag, or CSS to style content.

[Ntrails]

<div style="text-align:center; margin-bottom:10px;">Please use a <a href="/browsers">modern browser </a> with JavaScript enabled to use Coursera.</div>

Good job there folks

[rypskar]

At least they use noscript and give you a warning if you don’t have javascript enabled, there are so many sites that don’t check if javascript and cookies are enabled

[axansh]

Thanks coursera.

Thank you very much.

Looking forward for this course.

You save my some $$$ :) Will surely donate you.

[arek2]

I'll drop a link to my (non-free) e-book: http://arek-paterek.com/book/

[fmela]

So many great online courses, and so little time!