RFID数据流是一种动态RFID数据,它在某种因素的驱动下可能会随时间发生变化,而这种变化往往隐含着现实世界的某种事件.如何及时、准确地发现RFID数据流中的变化已成为数据流挖掘领域的一个研究热点,并且在实际中有非常广泛的应用.本文提出一种连续网格聚类算法来定位和描述RFID数据流的变化,并采用多分类器的方法建立变化特征向量与隐含事件的映射关系,实验表明该方法能有效检测出RFID数据流变化隐含的事件,与基于规则的事件检测方法相比,具有更高的检测效率.%RFID data stream is a kind of dynamic RFID data.It may change with time if it is influenced by some factors,and this kind of change always implies a certain event of the real world.How to find RFID changes in data streams timely and accurately has become a hot research field,and it has a very wide range of application in practice.This paper presents a continuous grid clustering algorithm to locate and describe the RFID data stream changes,and the use of multiple classifier method to establish the change feature vector and the mapping relationship between implicit event.Experiments show that this method can effectively detect the implicit event of RFID data stream changes,and it has higher detection efficiency compared with the rule based event detection method.
展开▼
