Radial and directional parts of a random vector
Let X be a random vector on and let R = [short parallel]X[short parallel] and for R [not equal to] 0 let W = W/R. Necessary and sufficient conditions are given for R and W to be independent. If X has a non-singular normal distribution we show that the following three conditions are equivalent. 1. (i) the components of X are independent and identically distributed with 0 means and positive variances. 2. (ii) W is uniformly distributed on the unit sphere. 3. (iii) R and W are independent.
Year of publication: |
1988
|
---|---|
Authors: | Jennrich, Robert I. ; Port, Sidney C. |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 6.1988, 3, p. 155-158
|
Publisher: |
Elsevier |
Keywords: | isotropic distributions normal distributions spherical distributions characterization of probability distributions |
Saved in:
Saved in favorites
Similar items by person
-
Brownian motion and classical potential theory
Port, Sidney C., (1978)
-
The Signed Root Deviance Profile and Confidence Intervals in Maximum Likelihood Analysis
Chen, Jian-Shen, (1996)
-
Diagnostics for Linearization Confidence Intervals in Nonlinear Regression
Chen, Jian-Shen, (1995)
- More ...