of different proposals to anomalous data. Several real data examples the multivariate exponentially weighted moving average procedure
are discussed. with the generalized likelihood ratio test based on nonparametric
The next two articles address problems in statistical process regression. It provides an effective SPC solution to handle nonlinear
control. In “The Null Steady-State Distribution of the CUSUM profiles and resolves the latent problem in popular parametric moni-
Statistic,” O. A. Grigg and D. J. Spiegelhalter develop an empiri- toring methods of being unable to detect certain types of changes
cal approximation for the cumulative sum (CUSUM) statistic. due to an incorrectly specified out-of-control model. Simulation
The derivation combines theoretical and empirical arguments, results demonstrate the effectiveness and efficiency of the monitor-
and the approximation is valid for CUSUMs applied to normal ing scheme. In addition, a systematic diagnostic approach is pro-
data with known variance (although the theoretical result is true vided to locate the change point of the process and to identify the
in general for exponential family data). The result leads to a for- type of change in the profile. Finally, a deep reactive ion etching
mula for steady-state p-values corresponding to CUSUM values, example from semiconductor manufacturing is used to illustrate the
where the steady-state p-value is obtained as the tail area of the null implementation of the monitoring and diagnostic approach.
steady-state distribution and represents the expected proportion of The final article, by Pritam Ranjan, Derek Bingham, and
time, under repeated application of the CUSUM to null data, that Michailidis is titled “Sequential Experiment Design for Contour
the CUSUM statistic is greater than some particular value. For Estimation from Complex Computer Codes.” Much of the recent
multiple CUSUM schemes, use of a p-value enables application of work on the design and analysis of computer experiments has
a signaling procedure that adopts a false discovery rate approach focused on scenarios where the goal is to fit a response surface or
to multiplicity control. The authors demonstrate their results optimize a process. In this article, the authors develop a sequential
by applying them to the number of earthquakes per year registering methodology for estimating a contour from a complex computer
> 7 on the Richter scale recorded from 1900 to 1998. code. The approach uses a stochastic process model as a surrogate
Industrial data used for process monitoring often take the form for the computer simulator. The surrogate model and associated
of nonlinear profiles. In their article, “Monitoring Profiles Based uncertainty are key components in a new criterion used to identify
on Nonparametric Regression Methods,” Changliang Zou, Fugee new computer trials aimed specifically at improving the contour
Tsung, and Zhaojun Wang propose a novel scheme for online mon- estimate. The authors apply their approach to exploration of a con-
itoring of such data that can detect changes in both the regression tour for a network queuing system. They also address issues related
relationship and the variation of the profile. The scheme integrates to practical implementation of their approach. n
L
A
T
E
X
L
A
T
E
X
L
A
T
E
X
MacKichan
SOFTWARE, INC.
NOVEMBER 2008 AMSTAT NEWS 11
AMSTAT November 08.indd 11 10/24/08 2:27:46 PM
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84