An architecture and methodologies for federated, privacy-enabled personalisation on the Web of Data

Users of the Social Web have come to expect personalised services for recommendations and prioritisation. At the same time, the awareness of users for privacy issues has increased. Yet, privacy and personalisation have conflicting objectives. Users need to make their profile data available in order to benefit from personalisation. Service providers on the other hand usually require access to the maximum available amount of data from the user. These developments require new methods and architectures for personalisation which takes federated sources, structured data and privacy into account. In this article we propose an architecture for federated, privacy-enabled eco-systems based on the WebID standard and the FOAF and Web Access Control vocabularies. It enables the creation of a universal private by default ecosystem which enables interoperability of user profile data while protecting the privacy of the user. In addition we describe two methodologies for providing personalisation on top of the proposed architecture and the Web of Data. First we describe and evaluate a methodology for using federated, structured data for multi-source recommendations. Then we describe a methodology for exploiting data from different topic domains for cross-domain recommendations. Combined, these two methodologies enable personalisation beyond the context of a single service, by taking user profile data into account from all sources of the users social graph as well as his interest graph.