A new variable sampling control scheme at fixed times for monitoring the process dispersion

Abstract

The variable sampling rate (VSR) schemes for detecting the shift in process mean have been extensively analyzed; however, adding the VSR feature to the control charts for monitoring process dispersion has not been thoroughly investigated. In this research, a novel VSR control scheme, sequential exponentially weighted moving average inverse normal transformation (EWMA INT) at fixed times chart (called (SEIFT) chart), which integrates the sequential EWMA scheme at fix times with the INT statistic, is proposed to detect both the increase and decrease in process dispersion. Moreover, the sample size at each sampling time is also allowed to vary. The Markov chain method is used to evaluate the performance of this new control chart. Numerical analysis reveals that this SEIFT chart gives significant improvement on detection ability than the fixed sampling rate schemes. Compared with other control schemes, the good properties of the INT statistic makes this SEIFT chart easy to design and convenient to implement.

Optimizing top precision performance measure of content-based image retrieval by learning similarity function

“…The experiments over some benchmark databases show its advantages over other similarity learning methods. In the future, we will investigate to use some other similarity function as similarity measure instead of linear function, such as Bayesian network [12], [13], [14], and also to develop novel algorithms of other machine learning problems and applications besides similarity learning, to maximize top precision measure, such as importance sampling [15], [16], [17], portfolio choices [18], [19], multimedia technology [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], computational biology [30], [31], [32], [33], [34], big data processing [35], [36], [37], [38], [39], computer vision [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], information security [54], [55], [56]…”

Adaptive Control Charts: Recent Developments and Extensions

“…Shi et al proposed a EWMA scheme based on the following inverse normal transformation statistic (IEWMA chart): where F n and Φ(·) are the cumulative distribution functions of the chi‐square distribution with n degrees of freedom and the standard normal distribution, respectively. The EWMA control statistic is then as the unweighted statistic, replacing with Z i .…”

“…The control statistic used in this scheme is similar to the unweighted EWMA control statistic presented in Reynolds and Arnold 60 replacing X k with ln s 2 k . Shi et al 65 proposed a EWMA scheme based on the following inverse normal transformation statistic (IEWMA chart):…”

Variable Sample Size and Sampling Interval Charts with Runs Rules for Switching Between Sample Sizes and Sampling Interval Lengths

“…See, for example, Amin and Widmaier, 6 Costa, 7 Aparasi and Haro, 8 Costa and Rahim, 9 Epprecht et al, 10 Zimmer et al, 11 Reynolds and Stoumbos, 12 Wu et al, 13 Yu and Hou, 14 Celano et al, 15 Chen, 16 Wu et al, 17 Yang and Su, 18 Mahadik and Shirke, 19,20 Jiang et al, 21 Jensen et al, 22 Zarandi et al, 23 and references therein. Some recent references include Luo et al, 24 Wu et al, 25 Shi et al, 26 De Magalhães et al, 27 Celano, 28 Faraz and Moghadam, 29 Mahadik and Shirke, 30,31 Li and Wang, 32 Dai et al, 33 and Epprecht et al 34 The drawback of any adaptive control chart is the inconvenience in its administration due to the frequent switches between the values of its adaptive design parameters. To reduce the frequency of switches between sampling interval lengths of VSI charts, Amin and Letsinger 36 proposed the use runs rules for switching between these lengths.…”