Benchmark · October 2, 2006 – September 21, 2009
Netflix announced an open contest to predict film ratings on its own dataset, with a cash award.
It entrenched the practice of public competition on a shared dataset under one metric — a format that later became routine for evaluating models.
The dataset was later criticised over the risk of re-identifying users, and a second contest was cancelled.
Netflix (archived) · Published 2009
Both belong to the period when model quality began to be compared on shared open data.
Both records are about an open contest over a dataset as a way to move a field.
The post describes which algorithms from the Netflix Prize's first Progress Prize the company put into production, and why it did not put in the Grand Prize ensemble.
Netflix Recommendations: Beyond the 5 stars (Part 1)The paper says that algorithms from the Netflix Prize contest still predict ratings in production.
The Netflix Recommender System: Algorithms, Business Value, and Innovation (ACM TMIS 6(4))