Matrix properties
number of rows	23,219
number of columns	23,219
nonzeros	325,592
structural full rank?	no
structural rank	8,210
# of blocks from dmperm	2
# strongly connected comp.	15,466
explicit zero entries	0
nonzero pattern symmetry	12%
numeric value symmetry	3%
type	integer
structure	unsymmetric
Cholesky candidate?	no
positive definite?	no

author	G. Kiss, C. Armstrong R. Milroy, J. Piper
editor	V. Batagelj
date	1971
kind	directed weighted graph
2D/3D problem?	no

Additional fields	size and type
nodename	full 23219-by-20

Notes:

------------------------------------------------------------------------------
Pajek network converted to sparse adjacency matrix for inclusion in UF sparse 
matrix collection, Tim Davis.  For Pajek datasets, See V. Batagelj & A. Mrvar,
http://vlado.fmf.uni-lj.si/pub/networks/data/.                                
------------------------------------------------------------------------------
 EAT - The Edinburgh Associative Thesaurus /                                  
 response-stimulus                                                            
 --------------------------------------------------------                     
 The EAT is a database of word association norms.                             
 - Original EAT: George Kiss, Christine Armstrong,                            
 Robert Milroy and J.R.I. Piper (1968-1971).                                  
 - MRC Psycholinguistic Database Version modified by:                         
 Max Coltheart, S. James, J. Ramshaw, B.M. Philip,                            
 B. Reid, J. Benyon-Tinker and E. Doctor;                                     
 made available by: Philip Quinlan.                                           
 - The present version was re-structured and documented                       
 by Michael Wilson at the Rutherford Appleton Laboratory.                     
 http://monkey.cis.rl.ac.uk/Eat/htdocs/eat.zip                                
                                                                              
 transformed in Pajek format: V. Batagelj, 31. July 2003                      
 -----                                                                        
------------------------------------------------------------------------------
Regarding conversion for UF sparse matrix collection: in the original data    
there are 325,624 weighted edges.  Of those only 32 edges are duplicates, and 
all of them have identical edge weights as the edges they are duplicates of   
These extraneous edges have been removed, since this this appears to be a     
graph, not a multigraph.                                                      
------------------------------------------------------------------------------

Ordering statistics:	AMD	METIS
nnz(chol(P*(A+A'+s*I)*P'))	23,400,769	29,261,922
Cholesky flop count	1.0e+11	1.5e+11
nnz(L+U), no partial pivoting	46,778,319	58,500,625
nnz(V) for QR, upper bound nnz(L) for LU	111,423,884	109,499,975
nnz(R) for QR, upper bound nnz(U) for LU	33,094,220	33,328,155

Maintained by Tim Davis, last updated 30-Sep-2008.
Matrix pictures by cspy, a MATLAB function in the CSparse package.
Matrix graphs by Yifan Hu, AT&T Labs Visualization Group.