c******a 发帖数: 45 | | c*******e 发帖数: 8624 | 2 什么叫怎么实现?
【在 c******a 的大作中提到】 : Thanks a lot!
| r****y 发帖数: 1437 | 3
Just numerically integrate it from x to a very large number, say 10000.
Depending on your requirement of precision, sometimes just to 100 would be
enough.
Or another approach, do integration from 0 to x, and you know the integral
from 0 to inf is sqrt(pi)/2. So you know the integral from x to inf.
Exp(-x^2) is dropping very fast from 0 to inf, accumulated contribution
outside 10 FWHM is really small.
【在 c******a 的大作中提到】 : Thanks a lot!
| w*********r 发帖数: 488 | 4 you can transform the variable x to its associated probaliliy given that you
know this varible has what kind of distribution. for example:
phi(x)=z, z belongs to [0,1], the unbounded integral is transformed into the
bounded integral ,but you also need to changed the pdf , new pdf is equal to
the "old" pdf times the Jacobian of the transformation.
【在 c******a 的大作中提到】 : Thanks a lot!
| b***n 发帖数: 53 | |
|