We exploit information about how people aggregate and organise musical artists and genres in
playlists to build a social-based model of multi-genre affinity for artists. In our model, inspired by
fuzzy sets theory, a membership degree indicates the affinity of an artist to a genre. This is contrast
to conventional labelling approaches, where artists either belong or do not belong to a genre.

We report on the analysis performed on a dataset of playlists compiled by the members of the
MusicStrands community, we present a richer ontology developed to describe artist-to-genre
associations, and show examples, domains, and applications where our work shows to be relevant.