The Research On Event Ontology Mapping Technology

As we know, there exist a large number of semantic conflict problems among heterogeneous data sources. It seems particularly important to solve the semantic heterogeneity while integrating data. The traditional ontology is mainly to describe concepts and relations between concepts, which causes ―Tennis Problem‖ easily. And, the traditional ontology in solving the problem of semantic heterogeneity is mainly through mapping the concepts to realize the integration of heterogeneous data, which can make data semantics easy to lose. By mapping the event classes, event elements as well as the event classes’ structure information, event ontology makes the event information semantic mapping to each other more abundant and the event elements mapping makes the accuracy of semantic mapping concept higher. For these heterogeneous data integration, event ontologies are extracted from the data source and semantic relations are discovered between events classes through the mapping of event ontologies, so as to realize the semantic understanding of the data source, and solve the problem of semantic heterogeneity.This paper mainly studies the event ontology mapping technology from different data source. First of all, we use the traditional ontology to complete the event ontology description, and construct a general event ontology representation model by reusing the traditional static concept. Then, we study the event ontology mapping of different data sources by using the unified event ontology representation model. The main research content and innovations are as follows:(1) Based on the related definition of the event, we extend the event, event elements and semantic of event classes. And we make a deep study on the relations between the event elements properties and relations between events and event elements. Furthermore, to make the event ontology semantic richer, the upper event classes and the sets of individuals are introduced to event ontology.(2) Based on the further research on the related concepts of events, we found the related static concepts like object, time, place elements in event play the different role, and have the corresponding properties. In the traditional ontology, the description of related concept, the relationship between concepts and the attributes of concept make it possible that event ontology reuse those concepts and their properties. According to the definition of related event elements and traditional ontology concepts, we adopt ontology mapping vocabulary in SKOS, to construct mapping relation with multiple traditional ontologies, which is described with OWL language, and finally construct a generic event representation model. The validity of the model is verified through a case.(3) First, the concept of event ontology mapping is proposed. On the basis, a comprehensive semantic similarity calculation model is proposed based on events and event classes’ structure information is proposed to map one event ontology to the other. While, the similarity of event classes is calculated by means of event sets, and the similarity of events is calculated based on event six elements. The semantic neighbors of event class are found in a certain semantic radius through event ontology network structure, which is further to find semantic information between event classes. The similarity of event classes’ structure information is obtained by calculating the similarity of semantic neighbor sets. By integrating the similarities of both, we finally obtain the final similarity between the event classes. Experiments show that the proposed model can effectively find the semantic relations of event classes in two event ontologies, which can solve the problem of semantic heterogeneity.