Regular D-optimal weighing designs with negative correlations of errors: new classes
Journal Title: Colloquium Biometricum; Colloquium Biometricum (Online) - Year 2017, Vol 47, Issue
Abstract
In this paper, the properties of the regular D-optimal chemical balance weighing design are considered. We consider this design under assumption that the measurements errors are equally negative correlated they have the same variances. Here we study the issues regard to the existence conditions of regular D-optimal design. We present the relations between the parameters of such design and construction methods.
Authors and Affiliations
Bronisław Ceranka, Małgorzata Graczyk
Keywords
Related Articles
- EP ID EP421262
- DOI -
- Views 107
- Downloads 0
MULTIDIMENSIONAL ASSESSMENT OF PHYSICOCHEMICAL PROPERTIES OF DIETARY BISCUITS
In the paper multivariate statistical techniques in combined analysis of measurements of physicochemical properties of dietary biscuits made according to different recipes have been applied. Relationships between two gro...
STRATUM ANALYSES FOR SPLIT-SPLIT-PLOT DESIGNS GENERATED BY GROUP DIVISIBLE DESIGNS
We present general and particular analyses under a mixed linear model resulting from the construction of a non-orthogonal split-split-plot (SSP) design and proper randomization performed (see, Ambroży and Mejza, 2013). T...
On the efficiency of orthogonal treatment contrasts in multistratum incomplete split-split-plot designs constructed by square lattice designs
We construct an incomplete split-split-plot design for three factor experiments. The method is based on a semi-Kronecker (Khatri-Rao) product of three matrices. We use two square lattice designs for whole plot treatments...
CHANGES OF CHOSEN PROPERTIES OF SANDY SOIL IN RESULT OF COMPOST APPLICATION
In this paper, the analysis of variance of the results of experiment conducted in laboratory conditions is described. The objective of this study was to evaluate changes in soil reaction, hydrolytic acidity, sum of bases...
Comparison of binomial proportions: new test
n the problem of comparison of two probabilities of success the most widely used is approximate test based on de Moivre-Laplace theorem. In the paper a test based on likelihood ratio is proposed. Those tests are compared...