Ontology Automatic Buiding And Mapping In Information Integration System

Along with the development and popularization of information integration and Internet, a large number of heterogeneous data sources came into being. In order to make better use of these resources, it is an urgent need to solve these heterogeneous data sources integration issues. Nowadays, ontology-based information integration becomes a research focus in information research domain.The paper proposes a new method to generate ontologies and mappings which are basis of information integration systems. The method constructs the ontologies and mappings from bottom-level to top-level which is different from the traditional building way. Local ontology and mapping between database and the local ontology are created at the bottom-level while global ontology and mapping between local ontology and global ontology at the top-level. In the local ontology building phase, this paper applies reverse engineering to extract semantic information and relational information from local data sources, and then builds the local ontology with the set rules. Meanwhile, the mapping between database and ontologly is generated by the information of building process of local ontology. In the global ontology building phase, a semantic dictionary is used to extract the semantics of the concepts in the local ontology. Then by calculating the semantic distance between the semantic meanings, the relationship between the concepts is generated. Finally, the local ontologies are merged into global ontology by the relations between concepts. The mapping between local ontology and global ontology is established by the semantic relation.In the building process, WordNet, a semantic dictionary, is utilized to obtain the semantics of concept. And then the similarity between concepts is figured out by the semantics. The final ontology is written by Protege OWL. At last, the method in this paper is verified by typical test cases.