| |
 |
|
|
Science Forum Index » Statistics - Math Forum » Statistical problem ...
Page 1 of 1
|
| Author |
Message |
| xlines |
 Posted: Sun Mar 11, 2007 10:25 am |
|
|
|
Guest
|
Hello!
I got statistical problem and any help is needed and appreciated =) Let say
I got a probability distribution in 3D vector space P(r), for simplicity let
say its constant on a ball of radius R. With that distribution, I can
calculate average values and variance of polar parameters of a vector. Now,
let say I take a sample of a n vectors ( where n is large, n ~ 10^5 ).
Problem is to calculate average distance between two vectors depending on
number n. My idea was to make approximation of :
Var (r) =~ n * (dr)^2
Where dr is average difference of parameter r of a two typical vectors among
n of them. Then, I would identify average distance between two vectors (
among n of them ) with distance of vectors ( <r>, <theta>, <phi> ) and
(<r>+dr, <theta>+dtheta, <phi>+dphi ). I am aware that this is incorrect
procedure, but is approximation valid in that regime ( n ~ 10^5 ). If not,
can you please point me how to calculate this correctly ? Thank you! |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Mon Dec 01, 2008 9:20 am
|
|