[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
## [igraph] thoughts on mixing matrices?

**From**: |
Bernie Hogan |

**Subject**: |
[igraph] thoughts on mixing matrices? |

**Date**: |
Thu, 27 Aug 2009 23:55:35 +0100 |

Hi various iGraph users,
I'm interested in calculating a mixing matrix for a graph, and then
using this matrix as a sort of 'reduced graph' for visualization. I
think of how to do it programmatically, but I'm wondering if such a
facility is available natively.
What is a mixing matrix? It is a graph where the rows and columns are
defined by some discrete values that are the properties of the nodes.
The values (either raw or normalized) refer to the number of links
between nodes having that that property. So a mixing matrix on
friendships between men and women would be a 2*2 matrix. The first
diagonal would be links between men, the second diagonal, links
between women. The off-diagonals are the men->women links and
women->men links. I noticed people mention assortativity here a little
while ago, and from what I understand, categorical assortativity can
be calculated by assessing the share of the diagonals versus the
off-diagonals. (Continuous assortitivity is more like a correlation
value, if I recall).
I have incredibly little experience programming in c, so I'd be
hesitant to add this myself.
Take care,
BERNiE
Bernie Hogan
Research Fellow, Oxford Internet Institute
University of Oxford

**[igraph] thoughts on mixing matrices?**,
*Bernie Hogan* **<=**