Authors: Xiaohua Cao; Xiejun Zhang
Addresses: School of Logisitics Engineering, Wuhan University of Technology, Wuhan City, 430063, China ' School of Logisitics Engineering, Wuhan University of Technology, Wuhan City, 430063, China
Abstract: In modern transportation logistics, anomaly significantly lowers the efficiency of production and the quality of service. Massive RFID data is produced to record the states of materials in transportation logistics. The data is of multi-attribute, randomness and various dimensions so that it is difficult to find out anomalies from these data. A deviation-based clustering approach is proposed to dig anomalies. Firstly, the features of RFID data are discussed from multi-attribute perspectives including time, location, data, sequence and path. Next, against the randomness and various dimensions of state data, a clustering approach is presented to unify the dimensions of state data and dig anomalies from random state data. The results show that the proposed approach can efficiently find more than 91.2% of anomalies among transportation logistics.
Keywords: anomaly digging; massive RFID data; radio frequency identification; deviation models; clustering; transport logistics; big data.
International Journal of Big Data Intelligence, 2014 Vol.1 No.3, pp.166 - 171
Received: 28 Nov 2013
Accepted: 10 Apr 2014
Published online: 16 Dec 2014 *