d*****o
发帖数: 61 | 1 two random normal x , y, x-N(0,sigma1), y-N(0, sigma2) , u=x+y,
What’s expectation of x conditional on u ?
What’s expectation of y conditional on u ?
Thanks! | d*****o
发帖数: 61 | 2 assume x and y independent | d*****o
发帖数: 61 | 3 assume x and y independent | F*******i
发帖数: 190 | 4 should both be u?
【在 d*****o 的大作中提到】
: two random normal x , y, x-N(0,sigma1), y-N(0, sigma2) , u=x+y,
: What’s expectation of x conditional on u ?
: What’s expectation of y conditional on u ?
: Thanks!
| d*****o
发帖数: 61 | | m******t
发帖数: 4077 | 6 no, it is linear estimator then. For Gaussian R.V., the best estimator is
the linear one. The formula is
E[X|Y] = E[X] + Cov(X, Y)Cov(Y, Y)^{-1}(Y -E[Y]).
【在 F*******i 的大作中提到】
: should both be u?
| f*********1
发帖数: 117 | 7 Agree.
no, it is linear estimator then. For Gaussian R.V., the best estimator is
the linear one. The formula is
E[X|Y] = E[X] + Cov(X, Y)Cov(Y, Y)^{-1}(Y -E[Y]).
【在 m******t 的大作中提到】
: no, it is linear estimator then. For Gaussian R.V., the best estimator is
: the linear one. The formula is
: E[X|Y] = E[X] + Cov(X, Y)Cov(Y, Y)^{-1}(Y -E[Y]).
|
|
