Ontology Based Trust Mechanism For The Semantic Web

Named graph is one of the solutions to the trust problem of the Semantic Web. The integrity and confidentiality of contents can be achieved by first issuing ontologies in the form of named graphs and specifying trust policies for the selection of demanded contents. To this aim, it is necessary to design trust policies and algorithms based on named graphs and the related technologies. Further, the algorithm should be implemented in a semantic web application.Trust policies and algorithms depend on named graphs, which are published through the approach of SWP (Semantic Web Publishing). Then RSWP (Semantic Web Publishing with Reputation) can be constructed on the basis of SWP by signing recommendations from other users. In this way, users can obtain the trustworthiness of contents even without the knowledge of their origin. The ontology-based trust algorithm, which is designed according to RSWP, consists of two steps as follows: search local information list about content-publisher, and choose the ontologies that they publish or recommend; then provide users with the ontologies which conform to the enforced constraints on trust-worthiness.Based on the algorithm mentioned above, we give a five-layer trust-based semantic web application framework: 1) Original Content layer, the set of local ontologies and foreign ontologies related to subjects; 2) Content Integration Layer, the selected trustworthy ontologies according to users'trust lists (ie. the first part of our algorithm); 3) Application Trust Constraint Layer, the constraints imposed on the selection of trustworthy ontologies (ie. the second part of our algorithm); 4)Query and Inference Layer, the mechanism of ontology inference, checking and results according to users'queries; 5)Result and Interpretation Layer, the procedure of responding users'queries with results and the inferred conclusions.