A framework for focused linked data crawler using context graphs

In this paper, we propose a framework for focused Linked Data (LD) crawler based on context graphs. A focused crawler searches for a specific subset of web, in our case it targets interlinked RDF data stores. The proposed crawler constructs set of context graphs for the given seed URIs by back crawling the web, and classifiers are trained to detect and assign documents to different categories based on the content type. These classifier help crawler in search and updating of context graphs automatically. The crawler are trained using supervised learning. Additionally, an extensive overview of existing LD crawlers is also provided along with its basic requirements, architecture, issues and challenges.