SPES Outlines JSM Program, Requests Volunteers
Michael Crotty, SAS Institute and SPES Awards Chair, and Stephanie P. DeHart, DuPont and SPES JSM Program Chair
Officers of the Section on Physical and Engineering Sciences (SPES) are seeking volunteers to help with the data collection for their awards program at JSM 2013. The awards, based on audience evaluations, are given for the best SPES-sponsored contributed and topic-contributed presentations and posters at JSM.
Volunteer responsibilities include distributing evaluation forms to the audience in each session and evaluating the presenters. Volunteers also are needed to visit the posters and provide evaluations.
SPES is the primary sponsor of three invited sessions, four topic-contributed sessions, six contributed sessions, one contributed poster session, and four roundtable discussions. The section also is cosponsoring several sessions of interest, including “A Celebration of J. Stuart Hunter’s Contributions to Technometrics and Statistics” on Tuesday at 2 p.m. Following are the details of the SPES-sponsored invited sessions. The complete online program is available on the JSM website.
Statistics in Chemistry and Chemical Biology
August 5, 10:30 a.m.
Organizer: Kirby Shedden, University of Michigan
This session will focus on statistical approaches to problems in chemistry and chemical biology. Major data analysis challenges in this field include management and analysis of very large data sets, accommodating complex sources of measurement error, multiple comparisons, extrapolation, and detection of weak signals. The session is aimed at applied statisticians interested in learning about applications of machine learning and data mining techniques to new areas, as well as to specialists in pharmaceutical sciences, toxicology, bioinformatics, chemical engineering, and related areas. Speakers include Shyamal Peddada, Arash Ali Amini, and Daniel Samarov.
Recent Advances in the Design of Multi-Stratum Experiments
August 6, 10:30 a.m.
Organizer: Peter Goos, University of Antwerp
Increasingly, many factorial experiments involve several stages of a production process. Complete randomization of the experiment is usually impossible. In this session, we discuss alternative experimental designs involving a restricted randomization (i.e., multi-stratum designs), as well as novel tools to set up restricted-randomization designs for several production steps. We pay attention to Hasse diagrams (a tool to visualize a design’s randomization structure), the combinatorial construction of regular fractional factorial two-level designs, the usefulness of integer linear programming in the construction of multi-stratum designs, and the importance of accounting for variance component estimation when designing multi-stratum experiments. All the talks will involve real-life examples of multi-stratum experiments in industry. Speakers include Randy Tobias, Eric Schoen, Heiko Grossmann, and Peter Goos.
Statistical Methods for the Greener Planet
August 7, 2:00 p.m.
Organizer: Ying Hung, Rutgers University
Green energy is extremely important for the future development of our globe. The study of green energy creates unique challenges and leads to new statistical issues. The purpose of this session is to bring three experts together to discuss the importance of efficient statistical methods in several cutting-edge green energy applications. Specific topics include statistical models for energy-efficient thermal management of data centers, uncertainty quantification in the study of building energy, and new methodologies from computer experiments for emulating and analyzing tidal power models. Beyond the community of industrial statistics, this session is expected to attract significant attention from those interested in statistics, computer science, and engineering. Speakers include Pritam Ranjan, Godfried Augenbroe, and Hendrik F. Hamann.