摘要：Intelligent security system includes inference system for context awareness and response. In wide areasurveillance system, many inference agents cooperate over exchanging knowledge. It requires efficientmanagement and searching of knowledge bases. It needs proper similarity measure technique to overcomethe difference in knowledge bases. Responding with the value of similarity within threshold helpscooperative inference. Various intelligent systems utilize ontology for scalability. Previous many similaritymeasure techniques uses static analysis mainly. In this paper ontology usage patterns are applieddynamically to improve existing method. Similarity measure is controlled by reflecting user's searchingpattern to edge weight. In performance evaluation the analysis results about effect of dynamic weightreflection to previous methods are shown.
关键词：similarity measure; ontology; intelligent surveillance; inference system
