and all 2048- and 4096-run designs of resolution 7. and Qualitative Input Variables.” This paper
A method is proposed for constructing minimum introduces a Bayesian methodology for prediction in
aberration designs using only a partial catalog of computer experiments having both quantitative and
good designs. Minimum aberration or good designs qualitative inputs. The proposed model is a hierar-
are tabulated up to 40, 80, 160, 45, 47, and 65 chical Bayesian model with conditional Gaussian
actors for 128, 256, 512, 1024, 2048, and 4096 stochastic process components. For each of the qual-
runs, respectively. itative inputs, the model assumes the outputs corre-
Our next article, by Gang Han, Thomas J. sponding to different levels of the qualitative input
Santner, and William I. Notz, studies a problem have ‘similar’ functional behavior in the quantitative
that arises in computer experiments: “Prediction inputs. The predictive accuracy of this method is
for Computer Experiments Having Quantitative compared with the predictive accuracies of alterna-
tive proposals in examples. The method is illustrated
in a biomechanical engineering application.
Curve estimation from observed noisy data has
broad applications. Jong-Hoon Joo and Peihua Qiu
Federal Committee on
consider this in “Jump Detection in a Regression
Curve and Its Derivative.” Their ideas are impor-
Statistical Methodology tant for applications in which the underlying
2009 Research Conference
regression curve may have singularities, including
jumps and roofs/valleys (i.e., jumps in the first
November 2–4, 2009
order derivative of the regression curve), at some
unknown positions, representing structural changes
The 2009 Federal Committee on Statistical Methodology of the related process. A number of jump-detection
(FCSM) Research Conference will be held November 2–4,
procedures have been proposed, most of which are
2009, at the Walter E. Washington Convention Center in
based on estimation of the (one-sided) first-order
Washington, DC.
derivatives of the true regression curve. Motivated
by related research in image processing, this article
The conference provides a forum for experts from around the proposes an alternative jump-detection procedure
world to discuss and exchange current research and method-
that exploits jumps in the second-order derivatives
ological topics relevant to federal government statistical pro-
and the first-order derivatives. Theoretical justifica-
grams. Each day of the conference will offer papers on a wide
tions and numerical studies show that this jump
range of topics, including the use of advanced technologies for
detector works well in applications. This procedure
is then extended for detecting roofs/valleys of the
survey design and data collection, processing and dissemina-
regression curve. A curve estimation procedure is
tion, data mining, data warehousing and metadata, treatment
also proposed, which can preserve possible jumps/
of missing data, improving coverage and response rates, confi-
roofs/valleys when removing noise.
dentiality and disclosure issues, record linkage, sample design
Much research in multiple regression has been
and estimation, cognitive research and usability testing, and devoted to identification of the best subset from
data quality. among a set of candidate predictors. In the case of
polynomial regression, the variable selection process
Technical demonstrations will run concurrently on the second
can be further complicated by the desire to obtain
day of the conference during the first morning session. Applica-
subsets that are hierarchically well-formulated, so
tions will include demonstrations of Q-Notes, Field Interviewer that inclusion of a high-order effect requires inclusion
(Fl) Tracker Tool, Statipedia, and census coverage measure- of lower-order effects in the same factors. Michael
ment clerical matching software.
J. Brusco, Douglas Steinley, and J. Dennis Cradit
develop an algorithm for computing such subsets in
Sessions will feature papers and demonstrations by govern-
their paper, “An Exact Algorithm for Hierarchically
ment, private sector, and academic researchers from eight Well-Formulated Subsets in Second-Order Polynomial
countries. Katharine Abraham from the Joint Program in Survey
Regression.” They present a branch-and-bound algo-
Methodology at the University of Maryland will be the guest
rithm for subset selection in second-order polyno-
speaker during the opening plenary session. All paper sessions
mial regression. They apply the new algorithm to
will include an open discussion; some will include a formal
a well-known data set from the regression literature
discussion.
and compare the results to those obtained from a
branch-and-bound algorithm that does not impose
Registration is $195. For a copy of the advance program and
the hierarchical constraints. The results of this
registration information, visit
www.fcsm.gov/events.
comparison reveal that the hierarchical constraints
yield only a small penalty in explained variation.
24 AmstAt News OCTOBER 2009
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 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100