THE EFFECT OF MISCLASSIFICATION DUE TO MEASUREMENT ERROR ON CUSUM CONTROL CHARTS FOR INTERVENED POISSON DISTRIBUTION

Abstract

DOI: 10.12957/cadest.2017.25564 In this paper the one-sided CUSUM chart for controlling the incidence and intervention parameters of the IPD under misclassification error due to measurement is discussed. Explicit formulae are derived for this purpose. The sensitivity of the parameters of the V-mask and the Average Run Length (ARL) is studied through numerical evaluation for grid of values. Numerical results presented reveal that the angle o of the mask increases slightly as shift in the ratio Ɵ Ɛ 1 / Ɵ Ɛ 2 decreases, whereas, for fixed α, the values of  decrease considerably as the deviation of Ɵ Ɛ0 from Ɵ Ɛ 1 increases. It is also shown that measurement error lessens the consumer’s risk, e 2 (because it gives early detection for the shift of the process parameter) and increases the producer’s risk, e 1 . Further for fixed ,e 1 , α, Ɵ Ɛ0 ,Ɵ Ɛ 1 , the values of ρ . Keywords: Measurement error, Misclassification, Intervened Poisson distribution.