Title: A method of cleaning RFID data streams based on Naive Bayes classifier

Authors: Qiao-min Lin; Yan Xiao; Ning Ye; Ru-chuan Wang

Addresses: College of Computer, Jiangsu High Technology Research Key Laboratory for Wireless Sensor Networks, Nanjing University of Posts and Telecommunications, Nanjing 210003, China ' Network Information Center, The Yangtze River Water Conservancy Committee, Nanjing 210011, China ' College of Computer, Jiangsu High Technology Research Key Laboratory for Wireless Sensor Networks, Nanjing University of Posts and Telecommunications, Nanjing 210003, China ' College of Computer, Jiangsu High Technology Research Key Laboratory for Wireless Sensor Networks, Nanjing University of Posts and Telecommunications, Nanjing 210003, China

Abstract: Recently, the radio frequency identification (RFID) technology has been widely used in many kinds of applications. However, RFID data streams contain false negative reads and false positive reads leading to the location uncertainty of RFID tags. In view of these problems, we propose a method of cleaning RFID data streams based on Naive Bayes classifier, which could detect effectively tags of false negative reads and false positive reads in RFID data streams. Firstly, we construct a model of a RFID data stream. Then we divide the method into three phases, i.e., preparation phase, training classifier phase and application phase. At last, the result of experiments illustrates our method based on Naive Bayes classifier could acquire the lower percentage of false negative reads and the higher percentage of false positive reads than SMURF algorithm with the increase of the size of sliding window.

Keywords: data stream cleaning; naive Bayes classifier; false negative reads; false positive reads; radio frequency identification; RFID data streams.

DOI: 10.1504/IJAHUC.2016.076359

International Journal of Ad Hoc and Ubiquitous Computing, 2016 Vol.21 No.4, pp.237 - 244

Received: 01 May 2014
Accepted: 16 Jan 2015
Published online: 05 May 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article