D-optimal designs via a cocktail algorithm
نویسندگان
چکیده
منابع مشابه
D-optimal designs via a cocktail algorithm
A fast new algorithm is proposed for numerical computation of (approximate) D-optimal designs. This cocktail algorithm extends the well-known vertex direction method (VDM; Fedorov 1972) and the multiplicative algorithm (Silvey, Titterington and Torsney, 1978), and shares their simplicity and monotonic convergence properties. Numerical examples show that the cocktail algorithm can lead to dramat...
متن کاملOn a Multiplicative Algorithm for Computing Bayesian D-optimal Designs
We use the minorization-maximization principle (Lange, Hunter and Yang 2000) to establish the monotonicity of a multiplicative algorithm for computing Bayesian D-optimal designs. This proves a conjecture of Dette, Pepelyshev and Zhigljavsky (2008). We also study a nearest neighbor exchange strategy and show that it can dramatically improve the speed of the multiplicative algorithm.
متن کاملNew D-optimal designs via cyclotomy and generalised cyclotomy
D-optimal designs are n x n ±l-matrices where n == 2 mod 4 with maximum determinant. D-optimal designs obtained via circulant matrices are equivalent to 2-{ v; kl i k2 i k1 + k2 ~(v 1)} supplementary difference sets, where v = ~. We use cyclotomy to construct D-optimal designs, where v is a prime. We give a generalisation of cyclotomy and extend the cyclotomic techniques which enables use to fi...
متن کاملSome new D-optimal designs
We construct several new (v; r, 8; A) supplementary difference sets with v odd and T' + .5 = A + (v 1) /2. They give rise to D-optimal designs of order 2v. D-optimal designs of orders 158, 194, and 290 are constructed here for the first time. We also give an up to date survey of this class of supplementary difference sets in arbitrary Abelian groups of odd order v < 100. o. Introduction Supplem...
متن کاملMinimax A- and D-optimal Integer-Valued Wavelet Designs for Estimation Abbreviated title: Minimax A- and D-optimal Wavelet Designs
In this study we discuss integer-valued designs for wavelet estimation of nonparametric response curves in the possible presence of heteroscedastic noise based on a modi ̄ed wavelet version of the Gasser-MÄ uller kernel estimator and weighted least squares estimation. The Gasser-MÄ uller estimator was modi ̄ed in order to obtain an exact expression for the bias of the estimator. We ̄rst use data s...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Statistics and Computing
سال: 2010
ISSN: 0960-3174,1573-1375
DOI: 10.1007/s11222-010-9183-2