Applications of number-theoretic methods to quantizers of elliptically contoured distributions
Kai‐Tai Fang, Ke‐Hai Yuan, Peter M. Bentler · Lecture notes-monograph series · 1994
In this paper we combine the so-called LBG algorithm in vector quantization and the number-theoretic methods to propose a new algorithm, NTLBG, of generating vector quantizer with low mean square error for many classes of multivariate distributions including one of the elliptically contoured distributions.Some numerical examples show that the present method is effective.* This paper is a small part of our presentation at The International Symposium on Multivariate Analysis and Its Applications.The primary contents of our presentation will be published in Statistical Science this year.