Método de agrupación de anotaciones en redes sociales (Clustering Method for Social Network Annotations)

Jose Javier Astrain (josej.astrain@unavarra.es), Francisco Echarte (patxi@eslomas.com), Alberto Córdoba (alberto.cordoba@unavarra.es), Jesús Villadangos (jesusv@unavarra.es)


Universidad Pública de Navarra
This paper appears in: Revista IEEE América Latina

Publication Date: March 2010
Volume: 8,   Issue: 1 
ISSN: 1548-0992


Abstract:
Folksonomies are a widely used tool of collaboratively creating and managing tags to annotate and categorize Internet resources (Web 2.0). The process of annotation and tag management by users of social networks is extremely easy and simple; however, it involves serious problems of navigation and search unlike what happens with taxonomies, thesauri and ontologies. The use of fuzzy similarity measures allows the correct identification of syntactic variations when tag lengths are greater or equal than five symbols, been inadequate for smaller length tags. This article presents a method that combines both fuzzy similarity and cosine measures in order to provide a proper classification of tags even with smaller tag lengths. This method allows the proper classification of the 95% of the syntactic variations of tags analyzed in the experiments.

Index Terms:
folksonomies, fuzzy similarity, social network, tag annotation, social tagging, tag clustering   


Documents that cite this document
This function is not implemented yet.


[PDF Full-Text (478)]