In the last years we have witnessed the rapid growth of a broad range of Semantic Web technologies that have been successfully employed to enhance information retrieval, data mining and user experience in real-world applications. Several authors have proposed approaches towards ontological user modelling in order to address different issues of personalized systems, such as the cold start problem. In all of these works, non-structured data such as tags are matched, by means of various techniques, against an ontology in order to identify concepts and connections between them. However, due to recent popularity of semantic metadata formats such as microformats and RDFa, structured data are often embedded in many Web contents, with no need to "guess" them using a support ontology which may not be coherent with the actual content and the original goals of the author. In this paper we propose a novel approach towards ephemeral Web personalization based on extraction and enrichment of semantic metadata embedded in Web pages. The proposed system builds, at client-side, a rdf network that can be queried by a content provider in order to address personalized content.
Users as crawlers: exploiting metadata embedded in Web pages for user profiling
DE NART, Dario;TASSO, Carlo;DEGL'INNOCENTI, Dante
2014-01-01
Abstract
In the last years we have witnessed the rapid growth of a broad range of Semantic Web technologies that have been successfully employed to enhance information retrieval, data mining and user experience in real-world applications. Several authors have proposed approaches towards ontological user modelling in order to address different issues of personalized systems, such as the cold start problem. In all of these works, non-structured data such as tags are matched, by means of various techniques, against an ontology in order to identify concepts and connections between them. However, due to recent popularity of semantic metadata formats such as microformats and RDFa, structured data are often embedded in many Web contents, with no need to "guess" them using a support ontology which may not be coherent with the actual content and the original goals of the author. In this paper we propose a novel approach towards ephemeral Web personalization based on extraction and enrichment of semantic metadata embedded in Web pages. The proposed system builds, at client-side, a rdf network that can be queried by a content provider in order to address personalized content.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.