Modelling Multimedia Social Networks Using Semantically Labelled Graphs

We live in an increasingly connected and data-greedy world. In the last decade, informative contents over the Web have grown in volume, connectivity and heterogeneity to an extent never seen before. Well known examples of Online Multimedia Social Networks (OMSNs), such as Facebook or Twitter, demonstrate the humongous volume and complexity characterizing common scenarios of the contemporary Web. Recognizing that, today, means adopting intelligent information systems able to use data and links between data to gain insights and clues from such intricate and dense networks. To address this goal, these systems should have formal models able to extract efficiently the knowledge retained in the network, even when it is not so explicit. In this way, complex data can be managed and used to perform new tasks and implement innovative functionalities. This article describes the use of a semantically labelled and property-based graph model in order to represent the information coming from OMSNs by exploiting linguistic-semantic properties between terms and the available low-level multimedia descriptors. The multimedia features are automatically extracted using algorithms based on MPEG-7 descriptors and integrated with textual data from a general knowledge base.