摘要
不确定上下文信息的表示与推理是研究和开发上下文感知系统的重点和难点。该文首次将证据理论和本体相结合,提出了基于证据理论的不确定性上下文本体建模方法,并对证据组合规则进行了修改,不仅解决了证据理论在高度证据冲突时的局限性,而且使得该推理模型具有自适应性,设计并实现了不确定上下文推理算法。最后,通过在原型系统中的医疗监护和诊断应用,验证了该方法的可行性、合理性和有效性。
The modeling and reasoning with uncertain context information is one of key problems of building context-aware applications in pervasive computing environment.In this paper,a method of modeling uncertain contexts is proposed by combining the D-S theory with ontology model.And also an improved evidences combination rule is presented not only to solve the D-S theory limitations in the case of high-degree conflict evidences but also to make the reasoning have self-adaptability.Then the uncertain reasoning algorithm is designed and implemented.Finally,the feasibility,rationality and effectiveness of this approach are verified through an e-health use case of the prototype system.
出处
《电子与信息学报》
EI
CSCD
北大核心
2010年第8期1806-1811,共6页
Journal of Electronics & Information Technology
基金
国家自然科学基金(60802034
60672122)
高等学校博士学科点专项科研基金(20070013026)
北京市科技新星计划(2008B50)资助课题
关键词
上下文建模
不确定性推理
证据理论
本体
Context modeling
Uncertain reasoning
Dempster-Shafter （D-S） theory
Ontology
