| The rapid development of the Semantic Web has greatly aroused computer scientists’enthusiasm on studying in ontology mapping. A lot of ontology mapping methods have been put forward. But still there is no single method can handle all tasks, often these perform well in some cases are not so well in some other cases. Therefore researchers have devoted to research on selecting the appropriate mapping method according to the specific task demand.At first, we introduce the research background of the subject, give a classification of ontology mapping technology, and introduce several ontology mapping systems which have achieved great accomplishment. Then we point out some developing directions of this field in the future.Secondly, we propose a mapping strategy selection method based on ontology features. After fully investigated the ontology features, we get the way to calculate5ontology features’characteristic value. And then we put forward a multi-strategy selection approach to select proper mapping methods for mapping task. For that purpose, we select several typical mapping methods corresponding to each feature as candidates, and make some improvements to some of them.Unlike the traditional ways to combine multi-strategies, this paper proposes a coordination degree based method to assign weights automatically to mapping strategies selected. This method does not need artificial participation, which improves the degree of automation of the mapping process. It can also overcome the weaknesses such as valuing all mapping strategies as the same or not considering all factors in traditional methods.Finally, in order to verify the superiority of the proposed multi-strategy selection and combination methods, this paper implements an ontology mapping system-FAMS(Feature-based Adaptive Mapping System). After doing experiments on it and comparing the results with several popular ontology mapping systems, it shows that the FAMS system can complete the mapping task with flying colors. |
PDF Full Text Request