We are pleased to announce the 2014 RecSys Challenge in conjunction with ACM RecSys 2014. The theme of this year's challenge is evaluation based on user engagement.
In the picture from left to right: Simon Dooms, Frédéric Guillou, Ernesto Diaz-Aviles, Alan Said, Róbert Pálovics, Domonkos Tikk.
The final leaderboard results can be found on the Leaderboard Page, the complete dataset as used for this challenge can be found on the MovieTweetings GitHub Page, full proceedings are available on the ACM Digital Library.
Evaluation is a key factor to reflect the quality of a recommender system algorithm. The theme of this year’s challenge is on evaluation. Instead of the typical rating prediction task, the goal is to predict how much interaction an item would receive via Twitter. The challenge is based on an extended version of the MovieTweetings dataset, which contains tweets that are generated when users rate movies on IMDb (using the iOS app). The challenge for participants is to rank such tweets by expected user interaction, which is expressed in terms of retweet and favorite counts.
A more detailed description of the challenge can be found in the Challenge section.
The RecSys challenge workshop will be held during the ACM RecSys 2014 at Foster City, Silicon Valley, USA, 6th-10th October 2014. Participants are expected to submit a paper detailing their results and findings. The winners of the challenge will be announced at the RecSys conference.