help > Symmetric connectivity matrix
Showing 1-4 of 4 posts
Sep 2, 2013 01:09 PM | David Coynel
Symmetric connectivity matrix
Hi,
When feeding a symmetric connectivity matrix into NBS, it seems to me that all possible connections are considered, e.g. N*(N-1)/2. Wouldn't it make more sense to only look at the upper triangular part of the matrix in this case ?
Thanks for your help,
David
When feeding a symmetric connectivity matrix into NBS, it seems to me that all possible connections are considered, e.g. N*(N-1)/2. Wouldn't it make more sense to only look at the upper triangular part of the matrix in this case ?
Thanks for your help,
David
Sep 2, 2013 01:09 PM | David Coynel
RE: Symmetric connectivity matrix
Woops, my bad. I was writing a bit too fast this time.
Sep 2, 2013 10:09 PM | Andrew Zalesky
RE: Symmetric connectivity matrix
Dear David,
Yes, you are absolutely right. Even though the user specifies a symmetric connectivity matrix, the software will only consider the upper triangular part of the matrix. In other words, only N*(N-1)/2 connections will be considered for statistical testing.
By extension, the current version of the software does not deal with asymmetric matrices (undirected graphs).
Andrew
Originally posted by David Coynel:
Yes, you are absolutely right. Even though the user specifies a symmetric connectivity matrix, the software will only consider the upper triangular part of the matrix. In other words, only N*(N-1)/2 connections will be considered for statistical testing.
By extension, the current version of the software does not deal with asymmetric matrices (undirected graphs).
Andrew
Originally posted by David Coynel:
Hi,
When feeding a symmetric connectivity matrix into NBS, it seems to me that all possible connections are considered, e.g. N*(N-1)/2. Wouldn't it make more sense to only look at the upper triangular part of the matrix in this case ?
Thanks for your help,
David
When feeding a symmetric connectivity matrix into NBS, it seems to me that all possible connections are considered, e.g. N*(N-1)/2. Wouldn't it make more sense to only look at the upper triangular part of the matrix in this case ?
Thanks for your help,
David
Sep 3, 2013 06:09 AM | David Coynel
RE: Symmetric connectivity matrix
That's what I meant. Thanks for the info.