Modeling Social Annotation: a Bayesian Approach - Computer Science > Artificial IntelligenceReportar como inadecuado

Modeling Social Annotation: a Bayesian Approach - Computer Science > Artificial Intelligence - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: Collaborative tagging systems, such as Delicious, CiteULike, and others,allow users to annotate resources, e.g., Web pages or scientific papers, withdescriptive labels called tags. The social annotations contributed by thousandsof users, can potentially be used to infer categorical knowledge, classifydocuments or recommend new relevant information. Traditional text inferencemethods do not make best use of social annotation, since they do not take intoaccount variations in individual users- perspectives and vocabulary. In aprevious work, we introduced a simple probabilistic model that takes interestsof individual annotators into account in order to find hidden topics ofannotated resources. Unfortunately, that approach had one major shortcoming:the number of topics and interests must be specified a priori. To address thisdrawback, we extend the model to a fully Bayesian framework, which offers a wayto automatically estimate these numbers. In particular, the model allows thenumber of interests and topics to change as suggested by the structure of thedata. We evaluate the proposed model in detail on the synthetic and real-worlddata by comparing its performance to Latent Dirichlet Allocation on the topicextraction task. For the latter evaluation, we apply the model to infer topicsof Web resources from social annotations obtained from Delicious in order todiscover new resources similar to a specified one. Our empirical resultsdemonstrate that the proposed model is a promising method for exploiting socialknowledge contained in user-generated annotations.

Autor: Anon Plangprasopchok, Kristina Lerman


Documentos relacionados