Collective Information Filtering for Web Observatories

Abstract

Web observatories are becoming a common on-line practice. Their role is to compile, organize and convey information that serves the needs of a thematically focused Web community. So far they are typically following a centralized approach, with an editorial team being responsible for finding, collecting, editing and presenting the observatory‘s information content. We propose a new approach for the development of Web observatories based on Collective Information Filtering. Community profiles are used to capture the collective interests of community members and evaluate the relevance of information content accordingly. We can thus build Web observatories that can be dynamically enriched and can continuously adapt their content to the interests/needs of the observatory‘s community. This new approach not only reduces significantly the cost of developing and maintaining a Web observatory, but also, following the current Web trends, it is community driven. In this chapter, we discuss Collective Information Filtering and we describe the architecture for applying it to a Web Observatory. We also present a series of prototype Web Observatories that adopt the proposed approach.