Limitation of RFID data cleaning method — SMURF

After the SMURF method was proposed in 2006, many researchers have found the inadequate of method; they proposed new method to improve this method. However, all work have been done focus on SMURF is less concern the mathematical theory. In this paper, we analyze the limitation of SMURF method on the view of mathematics; some results have been received by the analysis, the way to calculate parameter P_i^avg and the improper value of P_i^avg will both affect the evaluation of the size of smoothing windows and cause the decrease of SMURF performance. Based on the mathematical analysis, we present another confidence parameter η to decrease the false positive of original SMURF method. Meanwhile, we also provide some comment on the parameter δ, how to assign value of δ will affect the size of epoch numbers of smoothing window too. This will pave the way for an effective means of data warehousing system that will keep the RFID data safe for future mining.