Knowledge Discovery in Distributed Social Web Sharing Activities

Taking into consideration the steady shift towards information digitisation, an increasing number of approaches are targeting the unification of the user’s digital “Personal Information Sphere”to increase user awareness, provide singlepoint management, and enable context-driven recommendation. The Personal Information Sphere refers to both conventional information such as semi/structured information on the user’s personal devices and online accounts, but also in the form of more abstract personal information such as a user’s presence and activities. Online activities constitute a rich source for mining this type of personal information, since they are usually the only means by which a typical user consciously puts effort into sharing their activities. In view of this opportunity, we present an approach to extract implicit presence knowledge embedded in multiple streams of heterogeneous online posts. Semantic Web technologies are applied on top of syntactic analysis to extract and map entities onto a personal knowledge base, itself integrated within the wider context of the Semantic Web. For the purpose, we introduce the DLPO ontology—a concise ontology that captures all facets of dynamic personal information shared through online posts, as well their various derived links to personal and global semantic data clouds. Based on this conceptualisation, we outline the information extraction techniques targeted by our approach and present an as yet theoretical use-case to substantiate it.