The Effect of Introducing Content Price in Distributed Social Networks

Christos Tryfonopoulos
In Proceedings of the 23rd Conference on Information Technologies (IT), Kaunas, Lithuania, 2018

Over the last few years a number of distributed social networks with content management capabilities have been introduced both by academia and industry. However, none of these efforts has so far focused on supporting both information retrieval and filtering functionality in a distributed social net- working environment. In this work we present a social network- ing architecture that offers both functionalities –in addition to the usual social interaction tasks– in distributed social networks, outline the associated distributed protocols, and introduce a novel data source selection mechanism for identifying good data sources. This novel data source selection mechanism is designed to take into account a combination of resource selection, predicted publication behaviour, and content cost to improve the selection of information producers by users. To the best of our knowledge our approach, coined AGORA, is the first work to model the price of content and to study its effect on retrieval efficiency and effectiveness in a distributed social network setting. Finally, our work goes beyond modelling by providing proof-of-concept experiments with real-world corpora and social networking data.

Full text