Multi-stratum response surface designs

30 mins 45 secs, 107.70 MB, Windows Media Video 44100 Hz, 478.17 kbits/sec

Share this media item:

Embed this media item:

<iframe width="_width_" height="_height_" src="https://sms.cam.ac.uk/media/1428/embed" frameborder="0" scrolling="no" allowfullscreen></iframe>

Choose size:

About this item

Available Formats

About this item

Description:	Gilmour, S (QMUL) Monday 11 August 2008, 14:00-14:30

Created:

2008-09-09 06:28

Collection:

Design and Analysis of Experiments

Publisher:

Isaac Newton Institute

Gilmour, S

Credits:

Author:

Gilmour, S


Abstract:	Many industrial and laboratory-based experiments involve studying some factors which are hard, time-consuming or expensive to change and other factors which can be changed more easily. It is now well-recognised that varying some factors more quickly than others leads to split-plot or other multi-stratum structures. When the aims of the experiment lead to a desire to fit second-order polynomial response surface models, it is usually impossible to use a standard orthogonal design. Instead, an algorithm is usually used to search for a design which is optimal in some way. The different algorithms which have been suggested will be reviewed and the assumptions underlying their definitions of optimality will be reconsidered. The appropriateness of different criteria depend crucially on the objectives of the experiment and inappropriate scaling of the criteria can lead to misleading results. The differences and similarities between stratum-by-stratum and all-in-one methods of construction will be clarified. A Bayesian perspective will be used to show that, unless prior knowledge of the sizes of random effects is very certain or effects estimated in the higher strata are not of interest, experimenters should use either stratum-by-stratum methods or all-in-one methods with very large prior estimates of the higher stratum variance components. http://www.newton.ac.uk/programmes/DOE/seminars/index.html

Abstract:

Many industrial and laboratory-based experiments involve studying some factors which are hard, time-consuming or expensive to change and other factors which can be changed more easily. It is now well-recognised that varying some factors more quickly than others leads to split-plot or other multi-stratum structures. When the aims of the experiment lead to a desire to fit second-order polynomial response surface models, it is usually impossible to use a standard orthogonal design. Instead, an algorithm is usually used to search for a design which is optimal in some way. The different algorithms which have been suggested will be reviewed and the assumptions underlying their definitions of optimality will be reconsidered. The appropriateness of different criteria depend crucially on the objectives of the experiment and inappropriate scaling of the criteria can lead to misleading results. The differences and similarities between stratum-by-stratum and all-in-one methods of construction will be clarified. A Bayesian perspective will be used to show that, unless prior knowledge of the sizes of random effects is very certain or effects estimated in the higher strata are not of interest, experimenters should use either stratum-by-stratum methods or all-in-one methods with very large prior estimates of the higher stratum variance components.

http://www.newton.ac.uk/programmes/DOE/seminars/index.html

Available Formats

Format	Quality	Bitrate	Size
MPEG-4 Video	480x360	1.84 Mbits/sec	424.09 MB	View	Download
WebM	480x360	523.11 kbits/sec	117.50 MB	View	Download
Flash Video	480x360	805.62 kbits/sec	181.44 MB	View	Download
iPod Video	480x360	505.15 kbits/sec	113.77 MB	View	Download
QuickTime	384x288	848.05 kbits/sec	191.00 MB	View	Download
MP3	44100 Hz	125.03 kbits/sec	27.95 MB	Listen	Download
Windows Media Video *		478.17 kbits/sec	107.70 MB	View	Download
Auto	(Allows browser to choose a format it supports)