Computing the minimal efficiency of designs by a differentiable approximation of FEk-optimality criteria.docVIP

下载本文档

6
0
约2.03万字
约 12页
2017-05-08 发布于上海
举报
版权申诉

Computing the minimal efficiency of designs by a differentiable approximation of FEk-optimality criteria.doc

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Computing the minimal efficiency of designs by a differentiable approximation of FEk-optimality criteria

ActaMath.Univ.Comenianae Vol.LXXVII, 2(2008),pp.155–166 155 COMPUTING THE MINIMAL EFFICIENCY OF DESIGNS BY A DIFFERENTIABLE APPROXIMATION OF ΦEk-OPTIMALITY CRITERIA ˇ ′ L. BUSOVA Abstract. Consider the linear regression model with uncorrelated errors and an experimentaldesignξ. Inthepaper,weproposeanumericalmethodforcalculating theminimale?ciencyofξintheclassOoforthogonallyinvariantinformationcrite- (m) ria. Forthispurpose,weintroducetheconceptofΦk,p-optimalitycriteria. Thenwe showthatΦEk criteriacanbedi?erentiablyapproximatedbyΦ(k,pm) criteria,there- (m) fore it allows us to use standard numerical procedures to arrive at boundaries for (m) Φ optimalvalues,andhenceattheintervalsfortheminimale?ciencyofdesigns Ek under the class of all orthogonally invariant information criteria. The approach is thenillustratedonthepolynomial model ofdegrees2,...,8. 1. Introduction Theaimofthisarticleistonumericallycalculateboundaries fortheminimalef- ?ciencyofdesignswithrespect totheclassoforthogonallyinvariantinformation criteriaO,containingmanywell-knowncriteria,e.g. allKiefer’sΦ -optimalitycri- p teria(seee.g. [8,p.94]or[9,p.139]). Forthispurpose,itturnsouttobeuseful tostudypartialsumsofeigenvaluesofinformationmatrices. Asanexample,the sums of k smallest (or largest) eigenvalues can be used to characterize universal optimalityintheclassO,aswasshowninthearticleofBondar[1]. Thesumsof k smallesteigenvalues,viewedasspecialorthogonallyinvariantinformationfunc- tions Φ(m), are studied in [3]. The main result of the paper [3] is that minimal E k e?ciencyofagivendesignwithrespecttoOcanbecalculatedusingtheoptimal valuesofΦ(m)-criteria. Insomesimplermodels, theoptimalvaluescanbe calcu- E k latedexactly,withthehelpofequivalencetheoremforcriteriaofΦ(m)-optimality E k basedonspecial resultsfromconvexanalysis(see[7]). Moreover, boundaries for theoptimalvaluesofΦ(m)-criteriacanbecalculatedfromΦ -optimaldesigns(see E p [4]),buttheseboundariesar