Hi, I'm proud to announce that a paper called 'Scalable Similarity-Based Neighborhood Methods with MapReduce' written by me and my colleagues at Technische Universität Berlin was accepted for publication at the ACM Conference on Recommender Systems 2012.
The paper describes the algorithmic heart of Mahout's distributed recommender code, RowSimilarityJob, in detail and features experiments on various datasets, including one of the largest publicly available datasets from Yahoo Research, which consists of 700M ratings. http://recsys.acm.org/2012/program.html#top I will be at the conference and will be to meet anyone there interested in Mahout. Furthermore there will also be talk about recommendationg with Mahout at the RecSys Challenge workshop following the conference. Best, Sebastian
