Munich Personal RePEc Archive

From Marginals to Array Structure with the Shuttle Algorithm

Buzzigoli, Lucia and Giusti, Antonio (2006): From Marginals to Array Structure with the Shuttle Algorithm. Published in: Journal of Symbolic Data Analysis , Vol. 4, No. number 1 (June 2006): pp. 1-14.

[img]
Preview
PDF
MPRA_paper_49245.pdf

Download (316kB) | Preview

Abstract

In many statistical problems there is the need to analyze the structure of an unknown n-dimensional array given its marginal distributions. The usual method utilized to solve the problem is linear programming, which involves a large amount of computational time when the original array is large. Alternative solutions have been proposed in the literature, especially to find less time consuming algorithms. One of these is the shuttle algorithm introduced by Buzzigoli and Giusti [1] to calculate lower and upper bounds of the elements of an n-way array, starting from the complete set of its (n-1)-way marginals. The proposed algorithm, very easy to implement with a matrix language, shows interesting properties and possibilities of application. The paper presents the algorithm, analyses its properties and describes its disadvantages. It also suggests possible applications in some statistical fields and, in particular, in Symbolic Data Analysis and, finally, shows the results of some simulations on randomly generated arrays.

UB_LMU-Logo
MPRA is a RePEc service hosted by
the Munich University Library in Germany.