Permanent URL to this publication: http://dx.doi.org/10.5167/uzh21671
Barbour, A D; Chen, L H Y (2005). The permutation distribution of matrix correlation statistics. In: Barbour, A D; Chen, L H Y. Stein‘s method and applications. Singapore, 223245. ISBN 9812562818.

PDF (Preprint)
206kB 
Abstract
Many statistics used to test for association between pairs (yi, zi) of multivariate observations, sampled from n individuals in a population, are based on comparing the similarity aij of each pair (i, j) of individuals, as evidenced by the values yi and yj, with their similarity bij based on the values zi, and Zj. A common strategy is to compute the sample correlation between these two sets of values. The appropriate null hypothesis distribution is that derived by permuting the zi's at random among the individuals, while keeping the yi's fixed. In this paper, a Berry–Esseen bound for the normal approximation to this null distribution is derived, which is useful even when the matrices a and b are relatively sparse, as is the case in many applications. The proofs are based on constructing a suitable exchangeable pair, a technique at the heart of Stein's method.
Item Type:  Book Section, refereed, original work 

Communities & Collections:  07 Faculty of Science > Institute of Mathematics 
DDC:  510 Mathematics 
Language:  English 
Date:  2005 
Deposited On:  02 Feb 2010 19:16 
Last Modified:  17 Oct 2012 09:45 
Publisher:  World Scientific Publishing 
Series Name:  Lecture Notes Series. Institute for Mathematical Sciences. National University of Singapore. 
Number:  5 
ISSN:  17930758 
ISBN:  9812562818 
Additional Information:  Electronic version of an article published as Lecture Notes Series. Institute for Mathematical Sciences. National University of Singapore, 5. Published jointly by Singapore University Press, Singapore; and World Scientific Publishing Co. Pte. Ltd., Hackensack, NJ, 2005. xx+297 pp © 2005 copyright World Scientific Publishing Company http://www.worldscibooks.com/series/lnimsnus_series.shtml 
Official URL:  http://ebooks.worldscinet.com/ISBN/9789812567673/9789812567673_0015.html 
Citations:  Google Scholar™ 
Users (please log in): suggest update or correction for this item
Repository Staff Only: item control page