Etiquetagem de micromensagens no Twitter: uma abordagem linguística

• Ano de defesa: 2012
• Autor(a) principal: Evandro Landulfo Teixeira Paradela Cunha
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Folksonomias; Redes sociais online; Etiquetagem de conteúdo; Computação; Redes sociais on-line; Twitter; Indexação automatica
• Link de acesso: http://hdl.handle.net/1843/ESBF-8UZJ4X

Hashtags are labels used by Twitter members in order to classify messages posted in this social network. They are produced by the users themselves without any interference from the platform, which generates interest in studying them as linguistic elements since the appointment of a hashtag is driven by linguistic and social factors that inuence the creation of new tags and the acceptance of labels proposed by other members. In this work, we present a sociolinguistic-based study about the usage of hashtags on Twitter, assuming that its users' network has common features with oine speech communities, i.e., groups of people whose members linguistically inuence each other. Initially, we analyze the motivations that lead Twitter users to insert tags in their tweets. We found that the main reasons for labeling on Twitter are to increase the comprehensibility of the information and to raise the possibility of effective content sharing. Then, we examine some linguistic factors that contribute to success or failure of tags. Finally, we investigate the role of a social factor - the user's gender - in the usage of hashtags. Our results indicate that characteristics of some groups of hashtags are able to contribute to genderize them. The outcomes show similar features to those found in studies of oine speech, that leads us to believe that free tagging in folksonomies can serve as a model for characterizing the propagation of linguistic forms in other contexts. Our findings complement the knowledge about human behavior in free tagging environments and may be useful to increase the effectiveness of real-time streaming search algorithms and tag recommendation systems based on users' collective preferences.