摘要
为了解决在语义W eb信息处理中本体的频繁存取造成的性能问题,提出基于B ayes ian决策理论的本体缓存模型。该模型有效利用本体的语义性和本体的存取日志来抽取特征向量(包括语义特征、存取特征和类型特征),通过B ayes ian决策理论指导在本地缓存频繁使用的本体,并通过机器学习优化缓存模型,提高本体概念和实例缓存命中率。本体的有效缓存减少了本体网络访问的开销,实验表明,采用该本体缓存模型后,原型系统的本体访问速度在G auss分布的本体访问概率下提高了25%左右。
The performance of ontology-based information systems due that access large amounts of ontology data in distributed environments is improved with an ontology cached model based on Bayesian decision theory. The model reduces the network access frequency by locally and sequentially caching ontology data. The method abstracts several key characteristics from the ontology semantics and the ontology access history and then locally caches those ontology concepts or instances selected by the Bayesian decision algorithm. Tests show that the system increases the access rate by around 25%.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第10期1433-1435,1440,共4页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金资助项目(60443002)
