Welcome to our community

Be a part of something great, join today!

Correlation of Two Random Vectors

OhMyMarkov

Member
Mar 5, 2012
83
Hello everyone!

I'm coming to notice day by day how our education is purely focused on memorizing and applying formulas rather than understanding the concept. Assume we have the following:

$X = aR + N$, and
$Y = bG + W$,

where $X, Y$ are random vectors, $R, G$ are strongly correlated random vector that average out to the zero vector each, $a, b$ are scalars, and $N, W$ are two independent vectors of i.i.d. normal RVs.

Now, $X$ and $Y$ are correlated, right?
 

CaptainBlack

Well-known member
Jan 26, 2012
890
Hello everyone!

I'm coming to notice day by day how our education is purely focused on memorizing and applying formulas rather than understanding the concept. Assume we have the following:

$X = aR + N$, and
$Y = bG + W$,

where $X, Y$ are random vectors, $R, G$ are strongly correlated random vector that average out to the zero vector each, $a, b$ are scalars, and $N, W$ are two independent vectors of i.i.d. normal RVs.

Now, $X$ and $Y$ are correlated, right?
I somehow suspect you have missed out some information, but under my interpretation of what you mean, yes.

If you write out what you mean by correlation it should be obvious what the answer is.

CB

PS My interpretation of what you mean when you ask are X and Y correlated is that you are asking: is \( E( (X-\overline{X}) (Y-\overline{Y})^t)\ne {\bf{0}} \)?
 
Last edited:

OhMyMarkov

Member
Mar 5, 2012
83
Yes, this is what I mean. Wanted to make sure... I'll review the problem...