Structured data clouding across multiple webs

The variety of web resources available to users for their business or personal needs is growing, spanning from fast, short, ready-to-consume news/posts to well-structured, formal ontology instances of the Semantic Web. In this context, users require to retrieve very fast all available prominent information about target entities regarding events, people, situations and similar. In this paper, we introduce the notion of inCloud (information Cloud) and we propose an approach to web resource clouding for the construction of inClouds. inClouds are built for a target entity of interest by distinguishing, also in a visual way, how much prominent the retrieved web resource(s) are with respect to the target entity and by organizing web resources according to their reciprocal levels of closeness. An application of the proposed approach to a collection of real web resources about movies is presented. Applicability and evaluation issues are also discussed.

► We present data clouding as the activity of organizing webs data in form of inClouds. ► Data comes from multiple webs, including linked data, microdata and tags. ► An inCloud is built on a target entity according to data prominence and closeness. ► Techniques for closeness and prominence calculation are presented in the paper. ► Application issues and examples on a collection of real web resources are discussed.