m******e
发帖数: 45 | 1 How to generate uniform [0,1] distributed random numbers with correlation
rho?
generating correlated gaussian r.vs are trivial, but it seems there is no
clean way
for uniform r.v. , any idea? thanks. | c******s
发帖数: 90 | 2 Please see this link below.
http://www.noise.cz/sbra/sibram02/2-Ses/Fegan.htm
In this paper, the author present a way to generated correlated uniform R.V.
Another result you may be able to use is:
If X and Y are bivariate-normal with correlation rho,
Let Ux=normcdf(X) and Uy=normcdf(Y), then
Ux and Uy are bivariate-uniform with correlation
(6/pi)*arcsin(rho/2).
【在 m******e 的大作中提到】
: How to generate uniform [0,1] distributed random numbers with correlation
: rho?
: generating correlated gaussian r.vs are trivial, but it seems there is no
: clean way
: for uniform r.v. , any idea? thanks.
| m******e
发帖数: 45 | 3 consolas, Thanks very much! I will take a look of the link.
BTW, for the second method, is there any reference on how the correlation
(6/pi)*arcsin(rho/2) is obtained?
V.
【在 c******s 的大作中提到】
: Please see this link below.
: http://www.noise.cz/sbra/sibram02/2-Ses/Fegan.htm
: In this paper, the author present a way to generated correlated uniform R.V.
: Another result you may be able to use is:
: If X and Y are bivariate-normal with correlation rho,
: Let Ux=normcdf(X) and Uy=normcdf(Y), then
: Ux and Uy are bivariate-uniform with correlation
: (6/pi)*arcsin(rho/2).
| S*****x
发帖数: 20 | 4 I kind of remember Hull mentioned this in his book when he discussed credit
VaR. You can check it out.
【在 m******e 的大作中提到】
: How to generate uniform [0,1] distributed random numbers with correlation
: rho?
: generating correlated gaussian r.vs are trivial, but it seems there is no
: clean way
: for uniform r.v. , any idea? thanks.
| c******s
发帖数: 90 | 5 Essentially the correlated bi-variate uniforms generated
by the second method can be viewed as a Gaussian copula with marginal
uniform distribution. You can calculate the spearman's rho, which
is given by that formula.
【在 m******e 的大作中提到】
: consolas, Thanks very much! I will take a look of the link.
: BTW, for the second method, is there any reference on how the correlation
: (6/pi)*arcsin(rho/2) is obtained?
:
: V.
|
|
