Analyzing supersaturated designs by means of an information based criterion

Koukouvinos, C; Parpoula, C

dc.contributor.author	Koukouvinos, C	en
dc.contributor.author	Parpoula, C	en
dc.date.accessioned	2014-03-01T02:07:41Z
dc.date.available	2014-03-01T02:07:41Z
dc.date.issued	2012	en
dc.identifier.issn	03610918	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/29605
dc.subject	Error rates	en
dc.subject	Generalized linear models	en
dc.subject	Information theory	en
dc.subject	ROC	en
dc.subject	Screening	en
dc.subject	Supersaturated design	en
dc.subject	Symmetrical uncertainty	en
dc.subject.other	Error rate	en
dc.subject.other	Generalized linear model	en
dc.subject.other	ROC	en
dc.subject.other	Supersaturated designs	en
dc.subject.other	Symmetrical uncertainty	en
dc.subject.other	Data mining	en
dc.subject.other	Experiments	en
dc.subject.other	Information theory	en
dc.subject.other	Statistical mechanics	en
dc.subject.other	Uncertainty analysis	en
dc.subject.other	Design	en
dc.title	Analyzing supersaturated designs by means of an information based criterion	en
heal.type	journalArticle	en
heal.identifier.primary	10.1080/03610918.2011.579365	en
heal.identifier.secondary	http://dx.doi.org/10.1080/03610918.2011.579365	en
heal.publicationDate	2012	en
heal.abstract	The cost and time consumption of many industrial experimentations can be reduced using the class of supersaturated designs since this can be used for screening out the important factors from a large set of potentially active variables. A supersaturated design is a design for which there are fewer runs than effects to be estimated. Although there exists a wide study of construction methods for supersaturated designs, their analysis methods are yet in an early research stage. In this article, we propose a method for analyzing data using a correlation-based measure, named as symmetrical uncertainty. This method combines measures from the information theory field and is used as the main idea of variable selection algorithms developed in data mining. In this work, the symmetrical uncertainty is used from another viewpoint in order to determine more directly the important factors. The specific method enables us to use supersaturated designs for analyzing data of generalized linear models for a Bernoulli response. We evaluate our method by using some of the existing supersaturated designs, obtained according to methods proposed by Tang and Wu (1997) as well as by Koukouvinos et al. (2008). The comparison is performed by some simulating experiments and the Type I and Type II error rates are calculated. Additionally, Receiver Operating Characteristics (ROC) curves methodology is applied as an additional statistical tool for performance evaluation. Copyright © Taylor & Francis Group, LLC.	en
heal.journalName	Communications in Statistics: Simulation and Computation	en
dc.identifier.doi	10.1080/03610918.2011.579365	en
dc.identifier.volume	41	en
dc.identifier.issue	1	en
dc.identifier.spage	44	en
dc.identifier.epage	57	en