CiteULike is a user-article dataset, where each article has a 300-dimension tf-idf vector. XING is a user-view-job dataset where each job is described by a 2738-dimension...
The dataset used in the paper is a collection of implicit interactions gathered from various sources, including social networks and online communities.