Portal de Periódicos da CAPES

Inferring Semantic Facets of a Music Folksonomy with Wikipedia

2013; Routledge; Volume: 42; Issue: 4 Linguagem: Inglês

10.1080/09298215.2013.848904

1744-5027

Mohamed Sordo, Fabien Gouyon, Luís Sarmento, Òscar Celma, Xavier Serra,

Natural Language Processing Techniques

AbstractMusic folksonomies include both general and detailed descriptions of music, and are usually continuously updated. These are significant advantages over music taxonomies, which tend to be incomplete and inconsistent. However, music folksonomies have an inherent loose and open semantics, which hampers their use in many applications, such as structured music browsing and recommendation. In this paper, we present a system that can (1) automatically obtain a set of semantic facets underlying the folksonomy of the social music website Last.fm, and (2) categorize Last.fm tags with respect to the obtained facets. The semantic facets are anchored upon the structure of Wikipedia, a dynamic repository of universal knowledge.Keywords: information retrievalrepresentationsocial aspects AcknowledgmentsWe would like to thank reviewers of previous versions of this paper for their valuable comments that helped us improve the paper.Notes1 http://www.pandora.com2 The precise list of music attributes being a trade secret of the company. A now defunct Wikipedia page (Westergren, Citation2010) did provide until December 2012 a selection of such attributes.3 http://www.last.fm4 http://www.wordnet.princeton.edu5 http://www.last.fm/api6 http://www.last.fm/api/show/artist.getTopTags7 i.e. Last.fm ‘count’ or ‘relevance weight’, which goes from 0 to 1008 http://dbpedia.org9 http://www.w3.org/TR/2005/WD-swbp-skos-core-spec-2005110210 http://en.wikipedia.org/wiki/Wikipedia:SUBCAT11 People tend to tag an artist with the name of the artist.12 Except in the case of Mood for which our system does not retrieve any tag (see above). For this reason, we will not consider the Mood facet in evaluation scenario .13 This weight is given by the Last.fm dataset, and it is computed according to how many times tag has been applied to artist . The exact formula, though, is a trade secret of Last.fm.