Re: Distribution change using constant
- From: "Reef Fish" <large_nassua_grouper@xxxxxxxxx>
- Date: 21 Nov 2006 14:30:24 -0800
@ the Genius® @ wrote:
I have these normal distribution
X~N(-1,2)
Y~N(2,3)
X and Y are independent
Find the constant a1,a2,b1,b2 such that
a1*(X-b1)^2 + a2*(Y-b2)^2
If you take b1 = -1, and a1 = sqrt(1/2), Then (X-(-1))/sqrt(2) = Z1
Similarly, take b2 = 2, and a2 = sqrt(1/3), you have an independent Z2
and the sum of square of Z1 and Z2 would be your chisquare (2).
become a chi-square (2)
Find the constant c,d1,d2 such that
c(Y-d1) / |X-d2|
become a t-student (1)
Look at the requirement of T(1) as the quotion of a N(0,1)
and the square-root of an independent chi-square and a d.f.
factor, you can match c, d1, and d2 that way to fit what you
wanted.
-- Reef Fish Bob.
.
- References:
- Distribution change using constant
- From: @ the GeniusŪ @
- Distribution change using constant
- Prev by Date: Re: [Fwd: Mathematical Statistics]
- Next by Date: Re: (hyper)sensitivity of goodness-of-fit tests
- Previous by thread: Distribution change using constant
- Next by thread: Sum of two correlated Poisson distribution
- Index(es):
Relevant Pages
|
