16.11.2012 Views

'Advanced Network Analysis / Pajek: Large ... - Vladimir Batagelj

'Advanced Network Analysis / Pajek: Large ... - Vladimir Batagelj

'Advanced Network Analysis / Pajek: Large ... - Vladimir Batagelj

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

V. <strong>Batagelj</strong>: <strong>Analysis</strong> and visulization of large networks with <strong>Pajek</strong> 102<br />

✬<br />

Fast sparse matrix multiplication<br />

The standard matrix multiplication has the complexity O(|I| · |K| · |J|) – it<br />

is (usually) too slow to be used for large networks.<br />

For sparse large networks we can multiply faster considering only nonzero<br />

elements:<br />

for k in K do<br />

for i in NA(k) do<br />

for j in NB(k) do<br />

if ∃ci,j then ci,j := ci,j + ai,k ∗ bk,j<br />

else new ci,j := ai,k ∗ bk,j<br />

NA(k): neighbors of vertex k in network A<br />

NB(k): neighbors of vertex k in network B<br />

In general the multiplication of large sparse networks is a ’dangerous’<br />

operation since the result can ’explode’ – it is not sparse.<br />

✫<br />

Methodenforum der Fakultät für Sozialwissenschaften, Universität Wien, 21-22. 6. 2007 ▲<br />

❙ ▲<br />

▲<br />

▲<br />

● ❙ ▲<br />

▲<br />

☛ ✖<br />

✩<br />

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!