Organizer: Netflix, 2006-2009
Dataset: 100+ million historical user-movie ratings collected from Netflix users
Problem: automatically learn user preferences, in order to predict users’ ratings for unseen movies and provide most accurate personalized recommendations
Prize: $1 million for a rating prediction algorithm outperforming the in-house
Netflix system by 10%, as measured by RMSE of predictions
State of the art: Netflix’s Cinematch rolled out after several years of extensive research by Netflix R&D department
The most famous algorithmic competition ever
Before you click “I agree” or close the window, please read this information. We ask for your voluntary consent to the processing of your personal data by us and our business partners and provide information about the cookies notice and about the processing of your personal data by us. By clicking “I agree” or closing the window, you agree to the following. You can also refuse consent or limit its scope.
If you want to agree to processing of your personal data by TogetherData Sp. z o.o. that you provide in the history of browsing websites and web applications, for marketing purposes (including automated analysis of your activity on the website to determine your potential interests to adapt services, advertising or offer), including the placement of internet tags (cookies, etc. ) on your devices and reading such tags, click “I agree” or close this window.
Detailed information can be found in the