APPROXIMATE INDEPENDENCE OF DISTRIBUTIONS ON SPHERES AND THEIR STABILITY PROPERTIES
成果类型:
Article
署名作者:
RACHEV, ST; RUSCHENDORF, L
署名单位:
University of Munster
刊物名称:
ANNALS OF PROBABILITY
ISSN/ISSBN:
0091-1798
DOI:
10.1214/aop/1176990346
发表日期:
1991
页码:
1311-1337
关键词:
Extreme value theory
CONVERGENCE
rates
摘要:
Let zeta be chosen at random on the surface of the p-sphere in R(n), Op,n : = {x is-an-element-of R(n): SIGMA-i = 1n \x(i)\p = n}. If p = 2, then the first k components zeta-1,...,zeta-k are, for k fixed, in the limit as n --> infinity independent standard normal. Considering the general case p > 0, the same phenomenon appears with a distribution F(p) in an exponential class E. F(p) can be characterized by the distribution of quotients of sums, by conditional distributions and by a maximum entropy condition. These characterizations have some interesting stability properties. Some discrete versions of this problem and some applications to de Finetti-type theorems are discussed.