i*******g
发帖数: 13 | 1 Thanks for any reference/ help/ suggestion/ solution regarding the problem
as below:
If simulate η1, η2 η3,
where η1, η2 η3 are independent n(0, 1) distribution
we will get a sphere, right?
but now, if we constraint the simulated points only with 0.0003% joint
probability, simulate 1000 points, is that will be a surface of the sphere?
0.03% = Probability multivariate normal (η1,t≤ z1, η2,t≤ z2, η3,t≤ z3
)
is there a way that we can simulate the points with 0.0003% Joint
Probability, and | l*****i
发帖数: 3929 | 2 很好奇这个符号怎么打出来的
z3
【在 i*******g 的大作中提到】
: Thanks for any reference/ help/ suggestion/ solution regarding the problem
: as below:
: If simulate η1, η2 η3,
: where η1, η2 η3 are independent n(0, 1) distribution
: we will get a sphere, right?
: but now, if we constraint the simulated points only with 0.0003% joint
: probability, simulate 1000 points, is that will be a surface of the sphere?
: 0.03% = Probability multivariate normal (η1,t≤ z1, η2,t≤ z2, η3,t≤ z3
: )
: is there a way that we can simulate the points with 0.0003% Joint
| i*******g
发帖数: 13 | | k*****s
发帖数: 231 | 4 0.0003% Joint Probability can be simulated as it is. I guess the tricky
point in the problem is that since you simulated so many times (1000) with
so small joint probability, it should be the same as independent(0% joint
probability). So the simulation results are much the same. | A***l
发帖数: 302 | 5 Wonder whether importance sampling is what you are looking for.
z3
【在 i*******g 的大作中提到】
: Thanks for any reference/ help/ suggestion/ solution regarding the problem
: as below:
: If simulate η1, η2 η3,
: where η1, η2 η3 are independent n(0, 1) distribution
: we will get a sphere, right?
: but now, if we constraint the simulated points only with 0.0003% joint
: probability, simulate 1000 points, is that will be a surface of the sphere?
: 0.03% = Probability multivariate normal (η1,t≤ z1, η2,t≤ z2, η3,t≤ z3
: )
: is there a way that we can simulate the points with 0.0003% Joint
| i*******g
发帖数: 13 | 6 Thanks for all the help. !~ |
|
