000082 Katemee N;Mayureesawan T (Applied Statistics Dep, Faculty of Applied Science King, Mongkut's Univ of Technology, North B, Bangkok-10800, Thailand, Email: narun_ess@hotmail.com) : Cusum charts for zero-inflated generalized poisson process. Far East J mathl Sci 2013, 77(2), 225-43.
This paper develops four new cumulative sum charts (CUSUM charts) based on zero-inflated generalized Poisson (ZIGP) process. These charts can be used to detect shifts in the parameters for both the individual and simultaneity. The developed charts are as follows: the λz -CUSUM chart is used to detect the mean of nonconformities (λ) shifts, the ωz -CUSUM chart for detecting proportion of zero nonconformity (ω) shifts, the (ϕZ -CUSUM chart is used to detect over dispersion (ϕ) shifts and the TZ-CUSUM chart is for use in detecting simultaneous λ, ω and ϕ shifts. These developed charts are compared with the c-chart based on ZIGP distribution (cZ -chart). The performance was considered from the average run length (ARL). We have found that when there is simultaneous shift in λ, ω and ϕ, for some levels of parameters, the TZ -CUSUM chart is superior for low levels of shift. When there is only shift in λ for high levels of parameters, the λZ-CUSUM chart is the best for all levels of shifts. When there is only shift in ω for all levels of parameters, the ωZ -CUSUM chart performs for all levels of shifts. Finally, when there is only shift in ϕ for high levels of parameters, the ϕZ -CUSUM chart is the best for all levels of shifts.
4 illus, 2 tables, 12 ref