Re: multiple correlated random matrices
- From: "micromysore" <micromysore@xxxxxxxxx>
- Date: 26 May 2005 10:26:13 -0700
Hi,
When i meant correlated, i was referring to "corr2" in matlab.
Breaking the problem into 1D, 64x64 matrix can be referred to as a
vector of length 4096, letz says V = {n1,n2, ....,n4096};
so 2^4096, possible vectors, given the fact that n's are binary digits
{0's and 1's}, well 0's are mapped to -1's later
The message matrix of 8x8 can be broken down into
now let M = {n1,n2...n64};
so 2^64 possible messages, again n's are binary digits. {0's and 1's}
My problem wud be to find a mapping function ..
Given a Vector, V, and the message M, I should be able to generate
another vector V', i,.e. V' = F(V,M). the constraint here is that
corr(V,V') better be high.
High depends on what I trying achieve here .. In my case high sud be
such that ..
V1' = F(V1,M1)
k1 = corr(V1,V1')
V2' = F(V2,M1)
k2= corr(V1,V2')
(same message modulates different vectors V)
k1 > k2, for all V's
V1' = F(V,M1)
k1 = corr(V,V1')
V2' = F(V,M2)
k2= corr(V,V2')
(different messages modulate same vector V)
k1 (approx)= k2
-micromysore
.
- References:
- multiple correlated random matrices
- From: micromysore
- Re: multiple correlated random matrices
- From: Robert Israel
- multiple correlated random matrices
- Prev by Date: Re: Cantor and the binary tree
- Next by Date: Re: Cantor and the binary tree
- Previous by thread: Re: multiple correlated random matrices
- Next by thread: Any tricks to get variance of continuous data?
- Index(es):