Sentiment analysis in the planet art: A case study in the social semantic web

Abstract

Affective computing is receiving increasing attention in many sectors, ranging from advertisement to politics. Its application to the Planet Art, however, is quite at its beginning, especially for what concerns the visual arts. This work, set in a Social Semantic Web framework, explores the possibility of extracting rich, emotional semantic information from the tags freely associated to digitalized visual artworks, identifying the prevalent emotions that are captured by the tags. This is done by means of ArsEmotica, an application software that we developed and that combines an ontology of emotional concepts with available computational and sentiment lexicons. Besides having made possible the enrichment of the ontology with over four-hundred Italian terms, ArsEmotica is able to analyse the emotional semantics of a tagged artwork by working at different levels: not only it can compute a semantic value, captured by tags that can be directly associated to emotional concepts, but it can also compute the semantic value of tags that can be ascribed to emotional concepts only indirectly. The results of a user study, aimed at validating the outcomes of ArsEmotica, are reported and commented. They were obtained by involving the users of the same community which tagged the artworks. It is important to observe that the tagging activity was not performed with the aim of later applying some kind of Sentiment Analysis, but in a pure Web 2.0 approach, i.e. as a form of spontaneous annotation produced by the members of the community on one another's artworks.