Abstract
Recommender systems are popular social web tools, because they can address information overload and provide personalization of results rec05. This paper presents a collaborative approach to music playlist recommendation in a large scale scenario. We show that a simple markov-chain based algorithm improves performance compared to baseline models when neither content, nor user metadata is available. The lack of content-based features makes this task particularly hard.
Users
Please
log in to take part in the discussion (add own reviews or comments).