The Ontology Is Automatically Built Based On The Data Integration

As the development of Internet and widely using of informatin technology in all the fields,people are anxious to integrate all the heterogeneous data into one union forms which can be used conveniently.The aims of data integration is just for it.In the current,semantic heterogeneous is a big trouble for data intergration.Ontology is an explicit formal specification of how to represent the objects, concepts and other entities that are assumed to exist in some area of interest.It is important for resolving the semantic heterogeneous.In the current,most of the ontologies are built by the hands of the experts.There are some problems in it,such as complexity of the project,strongly depending on the experts,too slow for building ontology,the unclear terms.How to build ontologies automatically or semi-automatically,reducing the work of the experts and speeding up this proces is a hot-point for people.In this paper,we introduced the problems of data intergration and the relative knowledge about ontology at first.Also we sum up all the ways in data integration which based on ontology. Then we analyze the current problems in the process of ontology building.After the comparation among the three ways of automatically building ontologies,we bring out a new method to build ontology semi-automatically based on dictionary, which includes three stepps,concepts finding based on dictionary,relations finding based on dictionary and terms describing.For the data integration project of Information Port of CSU,we designed a system to build ontologies semi-automatically.It uses the technology of knowledge mining,proceeding through ontology import, extraction, prunning, refinement circularly,developping the domain ontology by the way of helix.The result of the experiment shows to us that this system can make a good function about the finding of new concept and the relations between the concepts,the description of the term good,and the ontology which is built aumatically has little difference to that which is modified by the experts.Finally, the paper sums up the whole research work and discusseshow to improve the system in the future.