Mixed Reasoning On RDF Data Sets From Multiple Fields

With the rise of big data and the Semantic Web, the data published on theweb is exploding, the efficient and reasonable use of data has become animportant topic of the Semantic application in various fields. In the dataintegration process, these vast amounts of data often get lost because of thescattered storage location, numbers and more artificial operation data errors.We need to deal effectively with data in order to use data efficiently, minemore knowledge in the field and find more valuable information. How toreason the data published on the Semantic Web, how to find implicitknowledge between information of the different fields, and how to improve thespeed of reasoning are currently the key technical problems we need to solve.It is necessary to reason using data in order to discover new knowledgesin the field. Reasoning technologies are mainly divided into three categories:ontology combined with description logic, ontology combined with rules andontology combined with inference engine. But these inferences are alwaysused in single field, and the most usage is forward inference of ontologycombined with rules or inference engine. The single filed inference involving a narrow field of information leads to discover less information. And reasoningin a single direction brings a low efficiency.In order to solve the problem that forward reasoning of data on singlefields not only makes the knowledge found limited but also makes reasoningprocess quite slow, this paper proposes a method of hybrid reasoning the dataof multiple fields, and constructs a mixed reasoning system in the field oftourism, transportation and finance using this method. First of all, we buildthree ontologies of tourism, transportation and financial using the ontologybuilding software Protege; Secondly, make the three ontologies alignment bybuilding SWRL rules in order to make the data sets associate in the fields oftourism, transportation and financial; Thirdly, build the mixing rules, andcreate a hybrid reasoning machine using Jena and Pellet reasoning machine;Finally, design a framework of the mixed reasoning system of data sets in thefields of tourism, transportation and financial, create a hybrid reasoning systemaccording to the framework, and do some contrast experiments. Afterexperiments, this method, compared to traditional ones, can supply moreinformation in a more efficient way and is practical.This paper designs a reasoning system based on the fields of tourism,transport and finance in the foundation of exploring the methods of hybridreasoning of data in multiple fields. This system could help users plan a tripdepending on different users’ travel needs. And it should be improved in theaspect of inference efficiency and algorithm optimization at present. The hybrid reasoning method in allusion to multiple areas of datasets in this papercan find more implicit knowledge than traditional method of forward inferencein the single filed. We hope it could be helpful to the researchers who isstudying ontology.