Database Reference
In-Depth Information
and the given utility matrix. The process converges to a local optimum, although to have a good chance of obtaining
a global optimum we must either repeat the process from many starting matrices, or search from the starting point in
many different ways.
✦
The NetFlix Challenge
: An important driver of research into recommendation systems was the NetFlix challenge. A
prize of $1,000,000 was offered for a contestant who could produce an algorithm that was 10% better than NetFlix's
own algorithm at predicting movie ratings by users. The prize was awarded in Sept., 2009.
9.7 References for Chapter 9
See [
5
] for a discussion of item-item similarity and how Amazon designed its
collaborative-filtering algorithm for product recommendations.
There are three papers describing the three algorithms that, in combination, won the
[1] G. Adomavicius and A. Tuzhilin, “Towards the next generation of recommender systems: a survey of the state-
of-the-art and possible extensions,”
IEEE Trans. on Data and Knowledge Engineering
17
:6, pp. 734-749, 2005.
[2] C. Anderson,
2004.
[3] C. Anderson,
The Long Tail: Why the Future of Business is Selling Less of More
, Hyperion Books, New York,
2006.
[4] Y. Koren, “The BellKor solution to the Netflix grand prize,”
2009.
[5] G. Linden, B. Smith, and J. York, “Amazon.com recommendations: item-to-item collaborative filtering,”
Internet
Computing
7
:1, pp. 76-80, 2003.
[6] M. Piotte and M. Chabbert, ”The Pragmatic Theory solution to the Netflix grand prize,”
2009.
[7] A. Toscher, M. Jahrer, and R. Bell, “The BigChaos solution to the Netflix grand prize,”
2009.