Performing Object Consolidation on the Semantic Web Data Graph

Abstract

An important aspect of Semantic Web technologies is the issue of identity and uniquely identifying resources, which is essential for integrating data across sources. Currently, there is poor agreement on the use of common URIs for the same instances across sources and as a result a naively in- tegrated dataset might miss associations between resources. To solve the problem, we present a method for performing large-scale object consolidation to merge identifiers of equivalent instances occurring across data sources; the algorithm is based on the analysis of defined inverse functional properties: properties which have values unique to an instance. We apply our object consolidation algorithm to a dataset of over 72M instances collected from more than 3M Web doc- uments, and o¿er evaluation and discussion on the resulting integrated graph.