Hugh Chipman, TechnometricsEditor

The February 2012 issue of Technometrics is filled with interesting articles. For a change, my summary consists of a “tweet” for each one, attempting to summarize the main points in 140 or fewer characters. Comments about this format are welcome at hugh.chipman@acadiau.ca —and need not be 140 or fewer characters.

“Massive Data Set Analysis for NASA’s Atmospheric Infrared Sounder” by Amy J. Braverman, Eric J. Fetzer, Brian H. Kahn, Evan M. Manning, Robert B. Oliphant, and Joao P. Teixeira
# Massive data reduction preserves distributional characteristics, enabling analysis.

“Bayesian Scale Space Analysis of Differences in Images” by Lasse Holmström and Leena Pasanen
# Smoothing at different levels tells us that some parts of these images are not like the others.

“Gaussian Process Single-Index Models as Emulators for Computer Experiments” by Robert B. Gramacy and Heng Lian
# New interpretation of “E (Y|x) = f(linear function of x),” with “f” a Gaussian process, for computer experiments

“Gaussian Surrogates for Computer Models with Time-Varying Inputs and Outputs” by Max D. Morris
# Weight function gives appropriate distance for Gaussian process covariance function with functional data.

“Robust Design Optimization with Quadratic Loss Derived from Gaussian Process Models” by Matthias H.Y. Tan and C.F. Jeff Wu
# Computationally tractable optimization in robust parameter design incorporates parameter uncertainty.

“Robust Designs for Poisson Regression Models” by J.M. McGree and J.A. Eccleston
# Optimal design of experiments when optimum depends on unknown parameters

“A Robust Standard Deviation Control Chart” by Marit Schoonhoven and Ronald J.M.M. Does
# Robust estimation in phase I protects against individual outliers and outlying groups.

“Nonparametric Estimation of a Periodic Sequence” by Ying Sun, Jeffrey D. Hart, and Marc G. Genton
# Leave-out-one cycle cross-validation finds integer periods in time series.