Skip to Content.
Sympa Menu

market-farming - [Market-farming] For Del mathematical and substantive meaning of the right

market-farming AT lists.ibiblio.org

Subject: Market Farming

List archive

Chronological Thread  
  • From: "Hook Family" <guldann AT ix.netcom.com>
  • To: <market-farming AT lists.ibiblio.org>
  • Subject: [Market-farming] For Del mathematical and substantive meaning of the right
  • Date: Tue, 11 Feb 2003 16:51:21 -0500

Ok list read no further if an OT, tongue in cheek email will bother you.
Note my name is Beth note the authors name of the following info :). Hint I
take credit for the Dna, but not the learn'

Hmm, well a correlation matrix is something from statistics. Let's say
you have a list of numbers. Call each number x_i where is just an index
so we have something like x_0 x_1 x_2 and so on. Now you can suppose that
these numbers are what we call random variables. Say X is a random
variable for tossing a coin. If the coin is fair we would have a
probability mass function p that said p(heads) = 1/2 and p(tails) = 1/2.
You can imagine the random variable being more complex and the function p
taking more arguments.

Now, imagine that you are flipping two coins at the same time. You would
have two random variables and the correlation between them is roughly how
much information you get about one coin flip from the other coin flip.
Since, coin flips don't effect each other, the correlation would be zero.

Now we could think of a different experiment. We are going to look at a
whole bunch of people. We have one random variable that is shoe size.
And we have another random variable that is...foot length. Now you know
what they say about guys who have big feet? They wear big shoes. But
there are still situations like clown shoes and kids whose feet are
growing faster than they can get new shoes. So the correlation is high,
but there is still some randomness involved.

Now for the matrix part. Remember that list? Well now we find the
correlation between each pair of elements in the list and we put it into a
box
cor(x_0, x_0) cor(x_1, x_0) cor(x_2, x_0) ...
cor(x_0, x_1) cor(x_1, x_1) cor(x_2, x_1) ...
and so on. In case you are wondering cor(x_1, x_0) = cor(x_0, x_1).

So now you have your correlation matrix. Eigenvectors are tricky things.
Vectors are lists of numbers and there are ways you can multiply vectors
with matrices. Say your matrix is M and your vector is V. Sometimes, it
happens that M*V = a*V where a is just a normal number. In this case V is
called an eigenvector and a is an eigenvalue. Eigenvectors and values are
useful for many things, well outside the scope of even this extended brief
response.

Chicken plucking?

Dan Hook






Archive powered by MHonArc 2.6.24.

Top of Page