This page contains a Flash digital edition of a book.
KEEP STORY AFTER CHANCE HIGHLIGHTS
Technometrics Highlights
Issue Takes on Future of
Statistical Computing
D
evelopments in the way we do statistics have been inti- microscopy and special software are used to identify and measure
mately linked to the way our computational environments nanotubes on a square grid. Current practice ignores nanotubes that
and tools have changed. As part of the 50th anniversary cross the grid lines, resulting in a length-biased sample. Paul Kvam
celebration of the journal, we asked Leland Wilkinson to give us his
shows how to correctly account for the biasing in his article, “Length
thoughts about the changes that lay ahead and how they will affect
Bias in the Measurements of Carbon Nanotubes.” He shows how
the practice of statistics. His article, “The Future of Statistical
to model the selection bias through an extended version of Buffon’s
Computing,” makes for fascinating and thought-provoking reading.
Needle Problem, constructs the nonparametric maximum likelihood
The article’s premise is that technology will influence statistical
estimator for the length distribution of the nanotubes, and examines
computing more than any other factor. The technology driving
the consequences of the length bias. Probability plots reveal that the
this forecast includes not only hardware, but also the software that
corrected length distribution estimate provides a better fit to the
provides the infrastructure for individual and community interac-
Weibull distribution than the original selection-biased observations,
tion with computers.
thus reinforcing a previous claim about the underlying distribution
Wilkinson writes that we are likely to see a proliferation of intel-
of synthesized nanotube lengths.
ligent data analysis systems embedded in everyday objects and web
The next article, “Measurement System Analysis for Binary
sites; automated visualizations for data discovery; analytic systems
Data” by Wessel N. van Wieringen and Jeroen de Mast, presents
that are accessible to nonstatisticians; distributed analytic systems
methods for assessing the quality of binary measurements, such as
that talk to each other, fuse disparate data in real time, and draw
pass-fail inspection systems. The authors thus extend the idea of
conclusions about the evidence; and communities of open-source
gauge repeatability and reproducibility (GR&R) from continuous
developers exceeding the scope and capabilities of commercial com-
to binary data. The article focuses on the situation where no refer-
panies. He predicts that statisticians interested in statistical comput-
ence values are available for the objects in the experiment. This
ing and its future incarnations will have to engage in joint research
leads the authors to use latent classes to model the results of the
with computer scientists to continue to have an influence.
R&R experiment. The paper provides maximum likelihood and
The article is accompanied by insightful discussions from John
method of moments estimators and compares their properties.
Chambers, Wayne Oldford, Doug Bates, Pat Hanrahan, Di Cook
Further, the paper gives guidelines for model-checking and recom-
with Hadley Wickham, and Duncan Temple Lang with Ross
mendations for sample sizes. The methodology is illustrated with
Ihaka. It closes with a rejoinder from Wilkinson.
an experiment from an industrial inspection process.
Research Articles
Johannes Forkman addresses another measurement problem
There are also eight research articles in the issue, covering an inter-
in his article, “A Method for Designing Nonlinear Univariate
esting array of scientific problems. The first concerns the difficult
Calibration.” He is concerned with designing a calibration experi-
and topical problem of using wireless sensor networks for monitor-
ment in which a set of standards of known value are measured
ing and tracking objects. Written by Natallia Katenka, Elizaveta
and there is a nonlinear relationship between the standards and
Levina, and George Michailidis, the article is titled “Robust Target
the measurements. He shows how to select a set of standards to
Localization from Binary Decisions in Wireless Sensor Networks.”
minimize the errors in the inverse predictions. The choice depends
The sensors have power and processing limitations so that the
on the curve parameters, and they are assumed to vary randomly,
information they collect needs to be appropriately fused before it is
between calibrations, with known expected value and known cova-
transmitted. The algorithms studied here are based on the local vote
riance matrix. A design criterion is suggested for analytical proce-
decision fusion (LVDF) mechanism, where sensors first correct their
dures, according to which the coefficient of variation and the area
original decisions using decisions of neighboring sensors. These cor-
under the precision profile are minimized.
rected decisions are more accurate, robust, and improve detection;
Graciela Boente and Andrés Farall are concerned with tol-
however, they are correlated, which renders maximum likelihood
erance regions. Their article, “Robust Multivariate Tolerance
estimation intractable. An extensive simulation study of the devel-
Regions: Influence Function and Monte Carlo Study,” defines a
oped algorithms, along with several benchmarks, establishes the over-
class of multivariate tolerance regions that are more resistant than
all superior performance of the LVDF-based algorithms, especially in
the classical ones to outliers. The tolerance factors are numerically
low signal-to-noise ratio environments. Extensions to tracking mov-
evaluated under the central model, and the sensitivity to devia-
ing targets and localizing multiple targets also are considered. tions from the normal distribution for moderate samples is studied
One of the hottest areas of current scientific research is nano- by Monte Carlo. The authors derive the influence function of the
technology. To measure carbon nanotube lengths, atomic force coverage probability, which allows them to compare the sensitivity
10 AMSTAT NEWS NOVEMBER 2008
AMSTAT November 08.indd 10 10/24/08 2:27:41 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
Produced with Yudu - www.yudu.com