7 Nov
2013
7 Nov
'13
12:11 p.m.
On 11/07/2013 08:01 AM, xenil wrote:
Two questions: - The usual katz calculation definition uses the transpose adjacency matrix, Is this transposition done in your code?
There are different conventions for the adjacency matrix of directed graphs. The code in the library follows the in-neighbours of the edge. So the matrix multiplication is performed: y_i = alpha \sum_j A_ij x_j + beta_i where A_ij is one if there is an edge in the direction j->i. If you want to transpose the matrix, it is easy: You just reverse the direction of the graph with g.set_reversed(True).
- Is the adjacency matrix normalized when no edge weight is used?
No. Cheers, Tiago -- Tiago de Paula Peixoto <tiago@skewed.de>