基于覆盖聚类的大本体分块映射研究

Research on large-scale ontology block mapping based on cover clustering

在本体映射的研究中,本体分块映射是完成大规模本体映射任务的主要方法。但是目前本体分块方法存在以下不足:分块大小不适合映射,分块质量依赖于预先给定的参数。针对这些问题,综合分块的内聚与耦合程度,作为分块指标指导分块过程,改善分块及映射的质量,并重新设计了基于覆盖聚类的本体分块映射的方法,使分块不依赖于参数。实验结果显示,改进的方法对大本体分块映射问题是行之有效的。

In the research of ontology mapping,ontology block mapping is the most popular of large-scale ontology mapping task.However,current partition methods have the following drawbacks：the size of block is not suitable for mapping,partition quality relies on the prior parameters.Aiming at solving those problems,this paper considers both cohesion and coupled degrees which guide the clustering process in order to improve the partition quality.Besides,this paper re-designs ontology block mapping method based on cover clustering with parameters unrelated.Experiment result shows that the method developed is fairly efficient for solving large-scale ontology mapping tasks.