D******n 发帖数: 2965 | 1 如何证明一个Gaussian Process的定积分是一个正态分布?我从网上查到物理系的一些
讲义里有这一结论,但数学上有无正规的定理或者结论?这里多谢了。
祝新年快乐! |
s***g 发帖数: 495 | 2 Is that a definition?
【在 D******n 的大作中提到】 : 如何证明一个Gaussian Process的定积分是一个正态分布?我从网上查到物理系的一些 : 讲义里有这一结论,但数学上有无正规的定理或者结论?这里多谢了。 : 祝新年快乐!
|
y***s 发帖数: 23 | 3 sum of normals is still normal (with different mean and variance)
The integral is the limit of finite sum (in the case of riemamn integral,
the other definition is ito-integral)
Hence we could expect the limit is still normal |
D******n 发帖数: 2965 | 4 如何证明一个Gaussian Process的定积分是一个正态分布?我从网上查到物理系的一些
讲义里有这一结论,但数学上有无正规的定理或者结论?这里多谢了。
祝新年快乐! |
s***g 发帖数: 495 | 5 Is that a definition?
【在 D******n 的大作中提到】 : 如何证明一个Gaussian Process的定积分是一个正态分布?我从网上查到物理系的一些 : 讲义里有这一结论,但数学上有无正规的定理或者结论?这里多谢了。 : 祝新年快乐!
|
y***s 发帖数: 23 | 6 sum of normals is still normal (with different mean and variance)
The integral is the limit of finite sum (in the case of riemamn integral,
the other definition is ito-integral)
Hence we could expect the limit is still normal |
D******n 发帖数: 2965 | 7 Thanks, 查到一些非正式的结果,也是这么说的。
【在 y***s 的大作中提到】 : sum of normals is still normal (with different mean and variance) : The integral is the limit of finite sum (in the case of riemamn integral, : the other definition is ito-integral) : Hence we could expect the limit is still normal
|