摘要
随着语义万维网(sematic Web)和关联数据集项目(linked data project)的不断发展,各领域的语义数据正在大规模扩增.同时,这些大规模语义数据之间存在着复杂的语义关联性,这些关联信息的挖掘对于研究者来说有着重要的意义.为解决传统推理引擎在进行大规模语义数据推理时存在的计算性能和可扩展性不足等问题,提出了一种基于Hadoop的语义大数据分布式推理框架,并且设计了相应的基于属性链(property chain)的原型推理系统来高效地发现海量语义数据中潜在的有价值的信息.实验主要关注于医疗和生命科学领域各本体之间的语义关联发现,实验结果表明,该推理系统取得了良好的性能———扩展性以及准确性.
With the development of Semantic Web and Linked Data Project,large amounts of semantic data from different kinds of domains is increasing rapidly.Meanwhile,there exist complex semantic associations among these data.The complex correlation information is of great importance for researchers.Conventional rule-based reasoning engines inevitably meet the bottleneck of computing performance and scalability when they are used to solve the problem of reasoning over big semantic data across large volumn ontologies.In this paper,we present a novel distributed reasoning framework for big semantic data based on Hadoop.Correspondingly,we also design a reasoning prototype system based on property chain to discover the implicit associations between different entities from the large-scale semantic data.In our experiment,we focus on the semantic association discovery in the areas of medical and life sciences.The experimental results indicate the prototype system achieves high performance,scalability and precision.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2013年第S2期103-113,共11页
Journal of Computer Research and Development