So I am faced with a challenge that I feel is easier than the method I am currently using to solve. I have an input matrix and a resulting output matrix.

The input matrix looks like this (just 400 x 500)

Parameter name number number number number number number number number
Parameter name number number number number number number number number
Parameter name number number number number number number number number
Parameter name number number number number number number number number
Parameter name number number number number number number number number

The out put matrix looks like this (just 5 by thousands)

Unitless Nuclide release rate release rate release rate
number name number number number
number name number number number
number name number number number
number name number number number
number name number number number

I am looking for a way to cross reference the first matrix with the second to find out which parameters in the first matrix are the most influential in the output located in the second matrix. Currently I have began learning Statistica to see if it will do what I want. Any suggestions?

6/30/2008 9:42:04 AM

Are you dealing with temporal data, spatial data, or data sampled from a population?

6/30/2008 12:01:00 PM

I'm working with a Total-System Performance Assessment (TPA) 5.1 package. The input matrix is made up of the parameter name and then the value assigned during each realization. Think of a realization sorta like a tally in MCNP. Each parameter has a range in which a random number is generated within that range for each realization. So the 400 x 501 matrix is a trial in which I ran 500 realizations. Hope this answers your question. If not let me know.

6/30/2008 3:32:30 PM

How do the dimensions of the input matrix relate to the dimensions of the output matrix?

[Edited on June 30, 2008 at 3:52 PM. Reason : clarity]

6/30/2008 3:51:58 PM

input is 400 x 501
output is 5 x 16002

7/1/2008 8:20:48 AM

If Statistica doesn't do what you need it to do, I know that SAS can do this in a very small amount of code... this is the kind of problem it handles best.

(disclaimer: i work for SAS, so i'm about as biased as you can get)

7/1/2008 9:49:23 PM

I'm downloading SAS right now. No one knows statistica. And the company won't offer support since it is not the most updated version.

7/9/2008 3:02:42 PM