Data Fusion Platform Based On Cross-Domain Ontology Linkage

In the era of rapid development of the information age,every moment every field has thousands of systems collecting data.When we solve practical problems in production activities,we often need to use these domains of data sets at the same time.The different information systems from different domains due to the way of building,forms of information storage,it would easily form a relatively independent information barriers.Traditional data fusion technology is only for different data sets in the same domain,but for cross-domain data set there's no mature method system.Therefore,the extraction and fusion of data from different domains is not only the urgent need of modern increasingly complex applications,but also the important technical challenges in the field of data fusion.Aiming at the above problems,this paper proposes a data fusion method based on cross-domain ontologies interconnection,and adopts ontology technology to solve the fusion problem between different domains.The fusion of data and information is supported by the integration of ontology concepts.As cross-domain ontologies have many differences in knowledge background,concept dimension and construction method,this paper proposes a matching method for cross-domain ontologies,builds multi-domain ontology network,and forms data sharing platform.The main contents of this paper are as follows:(1)A model of cross-domain ontology network is proposedIn this paper,the concept of cross domain ontology network is proposed for many independent ontologies existing on the Internet,and the concepts of different domain ontologies are collected,and these ontologies are organized graphically.The model has high versatility and expansibility.The ontology of conceptual intersection can be realized by this model.At the same time,the new domain ontology dynamic access is supported when the service is extended.(2)Designing ontology matching method according to the characteristics of cross-domain ontologiesDifferent ontologies are inconsistent in description language,knowledge background and concept dimension,so the traditional ontology matching method is not suitable for this type of matching demand.In this paper,an approach of "optimistic matching" is designed.Ontologies are firstly connected by text matching,and then more concepts and instance matches are found by using structural relations.(3)Using various domains to build corpora,and achieving semantic disambiguation from different dimensionsAccording to the different backgrounds of different fields,we construct domain corpus as the basis of semantic learning.At the same time,the method of multi-layer filtering is used to filter the conceptual matching of errors by using word disambiguation,structure disambiguation and associative items disambiguation to further improve the accuracy of ontology matching.(4)Through the establishment of ontology network,to achieve crossdomain data integrationBy constructing the ontology network,we can discover the knowledge representation of a certain concept in different fields,and extract the data of each part in a targeted way.After reassembling,the data is more fully described,which is the basis of data sharing and knowledge discovery.In this paper,we focus on cross-domain ontology network,ontology construction,ontology matching,word sense disambiguation and crossdomain ontology network construction,and design and validate the fusion method for cross-domain information characteristics.Through the integration of ontologies and the cross-domain data fusion,this paper proves that the proposed method has high universality and completeness for crossdomain data fusion,and can be applied to the increasingly complex crossdomain business needs.