Successive standardization of rectangular arrays (Q1736501)

Summary: In this note we illustrate and develop further with mathematics and examples, the work on successive standardization (or normalization) that is studied earlier by the same authors in [\textit{R. A. Olshen} and \textit{B. Rajaratnam}, Ann. Stat. 38, No. 3, 1638--1664 (2010; Zbl 1189.62109)] and [``Successive normalization of rectangular arrays'', in: Proceedings of the 1st international conference on data compression, communication and processing (CCP'11), Palinuro, Italy, 21--24 June 2011]. Thus, we deal with successive iterations applied to rectangular arrays of numbers, where to avoid technical difficulties an array has at least three rows and at least three columns. Without loss, an iteration begins with operations on columns: first subtract the mean of each column; then divide by its standard deviation. The iteration continues with the same two operations done successively for rows. These four operations applied in sequence completes one iteration. One then iterates again, and again, and again,\dots. In [the authors, 2010, loc. cit.] it was argued that if arrays are made up of real numbers, then the set for which convergence of these successive iterations fails has Lebesgue measure 0. The limiting array has row and column means 0, row and column standard deviations 1. A basic result on convergence given in [the authors, 2010, loc. cit.] is true, though the argument in [the authors, 2010, loc. cit.] is faulty. The result is stated in the form of a theorem here, and the argument for the theorem is correct. Moreover, many graphics given in [the authors, 2010, loc. cit.] suggest that except for a set of entries of any array with Lebesgue measure 0, convergence is very rapid, eventually exponentially fast in the number of iterations. Because we learned this set of rules from Bradley Efron, we call it ``Efron's algorithm''. More importantly, the rapidity of convergence is illustrated by numerical examples.