基于RFID路径数据的异常路径检测  

Mining abnormal path from RFID path data sets

作　　者：董国宾[1] 薛安荣[1] 赵保同[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《计算机应用研究》2013年第8期2451-2454,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(60773049);江苏省科技型企业创新基金资助项目(BC2010172);江苏大学高级专业人才科研启动基金资助项目(09JDG041);高校博士点基金资助项目(20093227110005)

摘　　要：RFID技术在物流、供应链管理等需要跟踪物品流动的领域的广泛使用产生了大量路径数据。路径数据描述了物品在RFID系统中的移动轨迹,路径数据中每个节点同时包含地点和时间信息,使路径数据比一般的序列数据更复杂。针对现有的序列数据异常点检测算法不适合处理路径数据的情况,对路径数据的异常路径检测进行了研究,提出适用于路径数据的扩展概率后缀树(EPST)模型和一种采用该模型检测异常路径的方法。该模型用来计算每个路径和路径数据集的相似度,在计算相似度时主要利用了路径数据的"短期记忆"性质,同时考虑了地点和时间信息对路径数据相似度的不同影响。实验表明提出的算法能够准确地检测出异常路径,并具有较低的空间复杂度。Radio frequency identification （RFID） technology is becoming a prevalent tool in tracking commodities in supply chain management application. The movement of commodities through the supply chain forms a gigantic path data. Each node in the path data contains location information and time information, so that the path data is more complex than the general se- quence of data. Existing sequence data outlier detection algorithms are not suitable for processing path data. According to the characteristics of the path data, this paper proposed extended probabilistic suffix tree （EPST） model to detect abnormal path. The similarity of a path and path data sets could be efficiently calculated with the EPST. The path data had a property of short memory and this could be used to simplify the calculation of similarity. The advantage of EPST was that when calculating the similarity it took into account the impact of the time and location information. The experiments show that the proposed algo- rithm can accurately detect the abnormal path and has a lower space complexity.

关 键 词：RFID 路径数据 异常路径检测 扩展概率后缀树 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]