论文信息 - Mixture models based on homogeneous polynomials

Mixture models based on homogeneous polynomials

Abstract Models for mixtures of ingredients are typically fitted by Scheffe’s canonical model forms. An alternative representation is discussed which offers attractive symmetries, compact notation and homogeneous model functions. It is based on the Kronecker algebra of vectors and matrices, used successfully in previous response surface work. These alternative polynomials are contrasted with those of Scheffe, and ideas of synergism and model reduction are connected together in both algebras. Scheffe’s ‘special cubic’ is shown to be sensible in both algebras.

Norman R. Draper | Friedrich Pukelsheim | F. Pukelsheim | N. Draper

[1] F. Pukelsheim,et al. On optimal third order rotatable designs , 1996 .

[2] F. Pukelsheim,et al. First and second order rotatability of experimental designs, moment matrices, and information surfaces , 1991 .

[3] Z. Galil,et al. Comparison of Simplex Designs for Quadratic Mixture Models , 1977 .

[4] H. Scheffé. The Simplex‐Centroid Design for Experiments with Mixtures , 1963 .

[5] H. Scheffé. Experiments with Mixtures , 1958 .

[6] Michael Jackson,et al. Optimal Design of Experiments , 1994 .

[7] F. Pukelsheim,et al. On third order rotatability , 1994 .

[8] M. Agha,et al. Experiments with Mixtures , 1992 .