A semantic-based approach for artist similarity

Oramas, Sergio; Sordo, Mohamed; Espinosa-Anke, Luis; Serra, Xavier

A semantic-based approach for artist similarity

License

Citation

Oramas S, Sordo M, Espinosa-Anke L, Serra X. A semantic-based approach for artist similarity. In: Müller M, Wiering F, editors. Proceedings of the 16th International Society for Music Information Retrieval (ISMIR) Conference; 2015 Oct 26 - Oct 30; Malaga, Spain. [S.l.]: International Society for Music Information Retrieval; 2015. p.100-6.

Abstract

This paper describes and evaluates a method for computing artist similarity from a set of artist biographies. The/nproposed method aims at leveraging semantic information present in these biographies, and can be divided in three/nmain steps, namely: (1) entity linking, i.e. detecting mentions to named entities in the text and linking them to an/nexternal knowledge base; (2) deriving a knowledge representation from these mentions in the form of a semantic/ngraph or a mapping to a vector-space model; and (3) computing semantic similarity between documents. We/ntest this approach on a corpus of 188 artist biographies and a slightly larger dataset of 2,336 artists, both gathered/nfrom Last.fm. The former is mapped to the MIREX Audio and Music Similarity evaluation dataset, so that its similarity/njudgments can be used as ground truth. For the latter dataset we use the similarity between artists as provided/nby the Last.fm API. Our evaluation results show that an approach that computes similarity over a graph of entitiesand semantic categories clearly outperforms a baseline that exploits word co-occurrences and latent factors.