Layered graph Data Model for dataspaces management

In order to improve management of heterogeneous data sources in dataspace, a unified data model to represent all kinds of data in a single, powerful and simple way is the foundation of dataspace system. We propose a layered graph Data Model (lgDM) namely entity data graph (G_D) and entity schema graph (G_S) to describe and capture associations among entities and entity classes. We propose an association mining strategy and a graph weighting scheme to facilitate keyword query and query results ranking. We conduct experiments to demonstrate the feasibility and functionality of our proposed data model.