Skip to content

Commit e1b01d6

Browse files
neoblizzneoblizz
committed
Example datasets.
1 parent 2fe01eb commit e1b01d6

File tree

1 file changed

+333
-0
lines changed

1 file changed

+333
-0
lines changed

‎datasets/chesapeake/chesapeake.mtx

Lines changed: 333 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,333 @@
1+
%%MatrixMarket matrix coordinate pattern symmetric
2+
%-------------------------------------------------------------------------------
3+
% UF Sparse Matrix Collection, Tim Davis
4+
% http://www.cise.ufl.edu/research/sparse/matrices/DIMACS10/chesapeake
5+
% name: DIMACS10/chesapeake
6+
% [DIMACS10 set: clustering/chesapeake]
7+
% id: 2457
8+
% date: 2011
9+
% author: unknown
10+
% ed: H. Meyerhenke
11+
% fields: name title A id date author ed kind notes
12+
% kind: undirected graph
13+
%-------------------------------------------------------------------------------
14+
% notes:
15+
% 10th DIMACS Implementation Challenge:
16+
%
17+
% http://www.cc.gatech.edu/dimacs10/index.shtml
18+
%
19+
% As stated on their main website (
20+
% http://dimacs.rutgers.edu/Challenges/ ), the "DIMACS Implementation
21+
% Challenges address questions of determining realistic algorithm
22+
% performance where worst case analysis is overly pessimistic and
23+
% probabilistic models are too unrealistic: experimentation can provide
24+
% guides to realistic algorithm performance where analysis fails."
25+
%
26+
% For the 10th DIMACS Implementation Challenge, the two related
27+
% problems of graph partitioning and graph clustering were chosen.
28+
% Graph partitioning and graph clustering are among the aforementioned
29+
% questions or problem areas where theoretical and practical results
30+
% deviate significantly from each other, so that experimental outcomes
31+
% are of particular interest.
32+
%
33+
% Problem Motivation
34+
%
35+
% Graph partitioning and graph clustering are ubiquitous subtasks in
36+
% many application areas. Generally speaking, both techniques aim at
37+
% the identification of vertex subsets with many internal and few
38+
% external edges. To name only a few, problems addressed by graph
39+
% partitioning and graph clustering algorithms are:
40+
%
41+
% * What are the communities within an (online) social network?
42+
% * How do I speed up a numerical simulation by mapping it
43+
% efficiently onto a parallel computer?
44+
% * How must components be organized on a computer chip such that
45+
% they can communicate efficiently with each other?
46+
% * What are the segments of a digital image?
47+
% * Which functions are certain genes (most likely) responsible
48+
% for?
49+
%
50+
% Challenge Goals
51+
%
52+
% * One goal of this Challenge is to create a reproducible picture
53+
% of the state-of-the-art in the area of graph partitioning
54+
% (GP) and graph clustering (GC) algorithms. To this end we
55+
% are identifying a standard set of benchmark instances and
56+
% generators.
57+
%
58+
% * Moreover, after initiating a discussion with the community, we
59+
% would like to establish the most appropriate problem
60+
% formulations and objective functions for a variety of
61+
% applications.
62+
%
63+
% * Another goal is to enable current researchers to compare their
64+
% codes with each other, in hopes of identifying the most
65+
% effective algorithmic innovations that have been proposed.
66+
%
67+
% * The final goal is to publish proceedings containing results
68+
% presented at the Challenge workshop, and a book containing
69+
% the best of the proceedings papers.
70+
%
71+
% Problems Addressed
72+
%
73+
% The precise problem formulations need to be established in the course
74+
% of the Challenge. The descriptions below serve as a starting point.
75+
%
76+
% * Graph partitioning:
77+
%
78+
% The most common formulation of the graph partitioning problem
79+
% for an undirected graph G = (V,E) asks for a division of V into
80+
% k pairwise disjoint subsets (partitions) such that all
81+
% partitions are of approximately equal size and the edge-cut,
82+
% i.e., the total number of edges having their incident nodes in
83+
% different subdomains, is minimized. The problem is known to be
84+
% NP-hard.
85+
%
86+
% * Graph clustering:
87+
%
88+
% Clustering is an important tool for investigating the
89+
% structural properties of data. Generally speaking, clustering
90+
% refers to the grouping of objects such that objects in the same
91+
% cluster are more similar to each other than to objects of
92+
% different clusters. The similarity measure depends on the
93+
% underlying application. Clustering graphs usually refers to the
94+
% identification of vertex subsets (clusters) that have
95+
% significantly more internal edges (to vertices of the same
96+
% cluster) than external ones (to vertices of another cluster).
97+
%
98+
% There are 10 data sets in the DIMACS10 collection:
99+
%
100+
% Kronecker: synthetic graphs from the Graph500 benchmark
101+
% dyn-frames: frames from a 2D dynamic simulation
102+
% Delaunay: Delaunay triangulations of random points in the plane
103+
% coauthor: citation and co-author networks
104+
% streets: real-world street networks
105+
% Walshaw: Chris Walshaw's graph partitioning archive
106+
% matrix: graphs from the UF collection (not added here)
107+
% random: random geometric graphs (random points in the unit square)
108+
% clustering: real-world graphs commonly used as benchmarks
109+
% numerical: graphs from numerical simulation
110+
%
111+
% Some of the graphs already exist in the UF Collection. In some cases,
112+
% the original graph is unsymmetric, with values, whereas the DIMACS
113+
% graph is the symmetrized pattern of A+A'. Rather than add duplicate
114+
% patterns to the UF Collection, a MATLAB script is provided at
115+
% http://www.cise.ufl.edu/research/sparse/dimacs10 which downloads
116+
% each matrix from the UF Collection via UFget, and then performs whatever
117+
% operation is required to convert the matrix to the DIMACS graph problem.
118+
% Also posted at that page is a MATLAB code (metis_graph) for reading the
119+
% DIMACS *.graph files into MATLAB.
120+
%
121+
%
122+
% clustering: Clustering Benchmarks
123+
%
124+
% These real-world graphs are often used as benchmarks in the graph
125+
% clustering and community detection communities. All but 4 of the 27
126+
% graphs already appear in the UF collection in other groups. The
127+
% DIMACS10 version is always symmetric, binary, and with zero-free
128+
% diagonal. The version in the UF collection may not have those
129+
% properties, but in those cases, if the pattern of the UF matrix
130+
% is symmetrized and the diagonal removed, the result is the DIMACS10
131+
% graph.
132+
%
133+
% DIMACS10 graph: new? UF matrix:
134+
% --------------- ---- -------------
135+
% clustering/adjnoun Newman/adjoun
136+
% clustering/as-22july06 Newman/as-22july06
137+
% clustering/astro-ph Newman/astro-ph
138+
% clustering/caidaRouterLevel * DIMACS10/caidaRouterLevel
139+
% clustering/celegans_metabolic Arenas/celegans_metabolic
140+
% clustering/celegansneural Newman/celegansneural
141+
% clustering/chesapeake * DIMACS10/chesapeake
142+
% clustering/cnr-2000 LAW/cnr-2000
143+
% clustering/cond-mat-2003 Newman/cond-mat-2003
144+
% clustering/cond-mat-2005 Newman/cond-mat-2005
145+
% clustering/cond-mat Newman/cond-mat
146+
% clustering/dolphins Newman/dolphins
147+
% clustering/email Arenas/email
148+
% clustering/eu-2005 LAW/eu-2005
149+
% clustering/football Newman/football
150+
% clustering/hep-th Newman/hep-th
151+
% clustering/in-2004 LAW/in-2004
152+
% clustering/jazz Arenas/jazz
153+
% clustering/karate Arenas/karate
154+
% clustering/lesmis Newman/lesmis
155+
% clustering/netscience Newman/netscience
156+
% clustering/PGPgiantcompo Arenas/PGPgiantcompo
157+
% clustering/polblogs Newman/polblogs
158+
% clustering/polbooks Newman/polbooks
159+
% clustering/power Newman/power
160+
% clustering/road_central * DIMACS10/road_central
161+
% clustering/road_usa * DIMACS10/road_usa
162+
%-------------------------------------------------------------------------------
163+
39 39 170
164+
7 1
165+
8 1
166+
11 1
167+
12 1
168+
13 1
169+
22 1
170+
23 1
171+
34 1
172+
35 1
173+
37 1
174+
39 1
175+
7 2
176+
8 2
177+
9 2
178+
11 2
179+
12 2
180+
13 2
181+
22 2
182+
23 2
183+
35 2
184+
36 2
185+
39 2
186+
14 3
187+
15 3
188+
16 3
189+
17 3
190+
18 3
191+
36 3
192+
39 3
193+
17 4
194+
36 4
195+
37 4
196+
39 4
197+
6 5
198+
34 5
199+
35 5
200+
39 5
201+
7 6
202+
35 6
203+
39 6
204+
8 7
205+
9 7
206+
11 7
207+
12 7
208+
13 7
209+
35 7
210+
39 7
211+
9 8
212+
10 8
213+
20 8
214+
21 8
215+
22 8
216+
23 8
217+
24 8
218+
35 8
219+
38 8
220+
39 8
221+
10 9
222+
35 9
223+
36 9
224+
39 9
225+
35 10
226+
36 10
227+
39 10
228+
19 11
229+
35 11
230+
36 11
231+
39 11
232+
19 12
233+
26 12
234+
35 12
235+
36 12
236+
38 12
237+
39 12
238+
35 13
239+
36 13
240+
38 13
241+
39 13
242+
25 14
243+
26 14
244+
27 14
245+
28 14
246+
29 14
247+
36 14
248+
39 14
249+
19 15
250+
25 15
251+
26 15
252+
27 15
253+
28 15
254+
29 15
255+
36 15
256+
39 15
257+
19 16
258+
27 16
259+
36 16
260+
39 16
261+
36 17
262+
39 17
263+
19 18
264+
25 18
265+
26 18
266+
27 18
267+
29 18
268+
32 18
269+
36 18
270+
39 18
271+
33 19
272+
36 19
273+
38 19
274+
39 19
275+
36 20
276+
38 20
277+
39 20
278+
33 21
279+
36 21
280+
38 21
281+
39 21
282+
27 22
283+
28 22
284+
30 22
285+
31 22
286+
32 22
287+
33 22
288+
35 22
289+
36 22
290+
38 22
291+
39 22
292+
30 23
293+
32 23
294+
33 23
295+
35 23
296+
36 23
297+
38 23
298+
39 23
299+
36 24
300+
38 24
301+
39 24
302+
36 25
303+
38 25
304+
39 25
305+
36 26
306+
38 26
307+
39 26
308+
30 27
309+
36 27
310+
38 27
311+
39 27
312+
36 28
313+
38 28
314+
39 28
315+
36 29
316+
38 29
317+
39 29
318+
36 30
319+
38 30
320+
39 30
321+
32 31
322+
36 31
323+
38 31
324+
39 31
325+
36 32
326+
38 32
327+
39 32
328+
36 33
329+
38 33
330+
39 33
331+
37 34
332+
36 35
333+
37 35

0 commit comments

Comments
 (0)