基于位置近邻的RFID路径聚类算法  

RFID path cluster algorithm based on location neighbor

作　　者：王霞[1] 赵龙 夏秀峰[2] 

机构地区：[1]沈阳职业技术学院计算机学院,沈阳110026 [2]沈阳航空航天大学计算机学院,沈阳110136

出　　处：《沈阳航空航天大学学报》2012年第2期46-50,共5页Journal of Shenyang Aerospace University

摘　　要：随着物联网技术的广泛应用,每天有大量RFID数据产生,如何从海量的位置数据信息中提取知识亟待解决。为了解决这一问题,考虑采用路径聚类的方法。为了确定路径之间的近邻关系,提出一种RFID位置模型LM(Location Model),利用该模型分析位置之间的相互联系,并提出一种RFID路径相似度计算算法,在此基础之上进行了路径聚类。经实验对比结果表明,基于位置近邻的RFID路径聚类算法可以更好的聚集相似路径,同时具有较低的时间复杂度和较高的准确率。There are increasingly more RFID data generated every day with the widening use of the technol- ogy of the Internet of things, and the problem is how to acquire knowledge from the massive location data. In order to determine the neighbor statement, this paper proposes a Location Model （LM） to compute the mutual relation of the locations. In addition, this paper presents a new algorithm to compute the similarities of the RFID paths, and cluster the paths based on the similarities. The experiment results show that the RFID path cluster algorithm based on location neighbor can cluster similar paths better, and has higher accu- racy as well as lower time complexity.

关 键 词：RFID数据 位置模型 路径相似度 路径聚类 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]