Sematic Association Discovery And Its Application From Linked Open Data

Semantic Web aims at attaching web pages with formal semantic information to make them machine understandable. With the rapid development of Semantic Web technologies, the effort of publishing, connecting structure data according to the principle of Linked Data will result in web of data. The establishment of Linked Linking Open Data community has greatly facilitated the release of the Linked Data. These published data sets cover a wide range of areas such as geographic information, demographic data and information, online communities, scientific publications, and music. The effort of publishing Linked Data provides potential for the discovery of hidden semantic association from data sources. As more and more Linked Data published, how to discover the semantic association from it become one of key issues in semantic web research field.Semantic association is defined as the representation of rich knowledge about binary relation in semantic data model .Semantic association Discovery is to infer out further associations based on existing semantic associations by inventing algorithm. However, existing methods are all based on a centralized data model, which is not consistent with the distributed features of the Linked Data, but also makes the scalability of the existing method poor. It's necessary to design a scalable semantic association discovery framework that is suitable for Linked Data.In response to these issues, we propose and implement a decentralized multi-agent collaborative framework for semantic association discovery. The main contribution is described as follow.①Present a new model of knowledge representation, introduced the hypothesis, evidence, evidentiary graph and other knowledge elements. The knowledge representation model will help the exchange of knowledge among the agents.②Propose a new multi-agent collaborative semantic association discovery mechanism.③Design and implemented two types of agents: Directory agents and worker agent, and provide detailed specifications for the services provided by the agents.④Design and implemented the core of semantic association discovery algorithms, research and analyze different strategies that could be employing in the algorithms.⑤Do simulation experiments to verify the feasibility of the framework and analyze the performance. We also apply the framework to DBLP and DBPedia data sets, the results show that it's feasible to use the framework to discovery the hidden semantic association.