CONICET | Buscador de Institutos y Recursos Humanos

Nowadays there are many on-line applications in which users? feedback is used to compute a trust value for products, services and also other users. In those application domains, trust models are required and extensively used. Since trust is context-dependent, currently, there is an increasing interest in the development of trust models that specifically consider the multi-context nature of trust. In this article, we propose a symbolic multi-context trust model where information will have a trust degree and an associated context. The set of available contexts will be related through a taxonomy of contexts that can be adapted to the application domain being modeled. We propose different alternatives to expand the information of one context using information from contexts that are related through the taxonomy. Algorithms for the implementation of our approach that consider strategies for handling conflicts will be introduced. Also, results of an empirical analysis of our proposal will be given.