Improving Cinemax, the Netflix Recommendation Engine
I’ve been fascinated with Netflix ever since I did a giant research paper about the company for my senior project for business management. Particularly fascinating to me were the technologies Netflix uses to estimate demand and the Netflix Recommendation Engine (Cinemax), which I have posted about previously.
Well now it looks like Netflix has placed a bounty on an algorithm to make the Recommendation Engine even better. My friend Peter Abilla indicates he might give it a try. He’s also posted about the rules and the $1 million dollar prize. I wish I had more time because this is one problem I would love to help solve. (I’m pretty nerdy like that.) Plus it would be cool to have a million bucks. :)
One thing I would recommend (kind of unrelated to the algorithm) is to allow users to rate movies with half, or even quarter starts. More precise input means better output, and I know I’ve had to round many times when I thought a movie deserved 3 and a half stars, etc.
Netflix has already unknowingly taken some of the advice my group suggested in the paper. We should have sent it Reed Hastings so we could claim credit for some of them. :) On a side note, if you have some free time and enjoy investigating companies and writing about them, there’s a pretty good opportunity for authorship at Startup Review. It’s a non-paid position, but you’d have an automatic audience. (Their readership numbers are pretty darn good.)
If you liked this post, you may want to subscribe to my RSS feed. Thanks for visiting!
Tags: Algorithm, Business, Netflix, Online Marketing, Peter Abilla, Programming, Recommendation Engine, Reed Hastings, Startup Review, Technology