forked from johannesgerer/jburkardt-cpp
-
Notifications
You must be signed in to change notification settings - Fork 0
/
lcvt_dataset.html
605 lines (556 loc) · 18.4 KB
/
lcvt_dataset.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
<html>
<head>
<title>
LCVT_DATASET - Latin Hypercubes using CVT Startup
</title>
</head>
<body bgcolor="#EEEEEE" link="#CC0000" alink="#FF3300" vlink="#000055">
<h1 align = "center">
LCVT_DATASET <br> Latin Hypercubes using CVT Startup
</h1>
<hr>
<p>
<b>LCVT_DATASET</b>
is a C++ program which
computes a Latin Hypercube in M dimensions,
with N points, using a CVT dataset as the initial estimate.
The resulting dataset can be written to a file.
</p>
<p>
A Latin Square dataset is typically a two dimensional dataset
of <b>N</b> points in the unit square, with the property that, if both the
<b>x</b> and <b>y</b> axes are divided up into <b>N</b> equal subintervals,
exactly one dataset point has an <b>x</b> or <b>y</b> coordinate in
each subinterval. Latin squares can easily be extended to the
case of <b>M</b> dimensions, and may be pedantically called <i>Latin
Hypersquares</i> or <i>Latin Hypercubes</i> in such a case.
Statisticians like Latin Squares, as
do experiment designers, and and people who need to approximate
scalar functions of many variables.
</p>
<p>
The fact that the projection of a Latin Square dataset onto any
coordinate axis is either exactly evenly spaced, or approximately
so (depending on the algorithm), turns out to be an attractive
feature for many uses.
</p>
<p>
However, a CVT dataset in a regular domain, such as the unit
hypercube, has the tendency for the projections of the points
to cluster together in any coordinate axis. This program is
mainly an attempt to explore whether a dataset can be computed
using techniques similar to those of a CVT, but with the
constraint (whether imposed or expected) that the point projections
do not clump up.
</p>
<p>
The approach used here is quite simple. First we compute a CVT
in M dimensions, comprising N points. We assume that the bounding
region is the unit hypercube. We are now going to adjust the
coordinates of the points to achieve the Latin Hypercube property.
For each coordinate direction, we simply sort the points by that
coordinate, and then overwrite the original values by the values
we'd expect to get for a centered Latin Hypercube, namely,
1/(2*N), 3/(2*N), ..., (2*N-1)/(2*N).
</p>
<p>
Now this process guarantees that we get a Latin Hypercube. Our
hope is that the process of adjusting the point coordinates does
not too severely damage the nice dispersion properties inherent
in the CVT point placement.
</p>
<p>
An earlier version of this program was "very" interactive,
allowing the user to enter input in any order. This turned
out to be a little too confusing. The new version of the
program asks the user for input in a strict order. If you
find this procedure too restrictive, you can try out the
old program.
</p>
<p>
Briefly the user needs to specify the following:
<ol>
<li>
The spatial dimension M of the points;
</li>
<li>
The number of points N to be generated.
</li>
<li>
The random number seed;
</li>
<li>
How the initial points are chosen. If you have no preference,
choose UNIFORM.
<ul>
<li>
GRID, use a grid of points;
</li>
<li>
HALTON, use Halton points;
</li>
<li>
RANDOM, use RAND (C++ intrinsic);
</li>
<li>
UNIFORM, use a simple uniform random number generator;
</li>
<li>
USER, call the "user" routine;
</li>
<li>
(file_name), read the initial points from a file.
</li>
</ul>
</li>
<li>
The number of CVT iterations. If you have no preference,
try 5, 10 or 20;
</li>
<li>
How the sampling is done. If you have no preference, use UNIFORM.
<ul>
<li>
GRID, use a grid of points;
</li>
<li>
HALTON, use Halton points;
</li>
<li>
RANDOM, use RAND (C++ intrinsic);
</li>
<li>
UNIFORM, use a simple uniform random number generator;
</li>
<li>
USER, call the "user" routine;
</li>
</ul>
</li>
<li>
The number of sampling points to use. Think of this as
a sampling of the unit hypercube. So to compare it to
N, the number of points, you need to take its M-th root.
In 2D, if you're using 10 generators, and 100 sample points,
to get area and sampling computations twice as good requires
4 times the sampling. It never hurts to use more sampling
points.
</li>
<li>
The "batch size". This parameter controls how many sampling
points are to be generated at one time. You can set this
value equal to the number of sampling points, but if you
are having memory problems, it can be set lower. In such
a case, a smaller value might be 1000, for instance.
</li>
<li>
The number of CVT iterations to carry out. It's not really
necessary to compute the CVT super accurately, since we're
just going to perturb it anyway. This value could be anywhere
from 10 to 500. Convergence of the CVT is typically slow,
especially if the starting positions are poor.
</li>
<li>
The number of Latin Hypercube iterations to carry out.
Actually, the iterations don't seem to improve the data
much, so a value of 1 or 2 can be reasonable.
</li>
<li>
The name of a file into which the final pointset should be
written.
</li>
</ol>
</p>
<h3 align = "center">
Licensing:
</h3>
<p>
The computer code and data files described and made available on this web page
are distributed under
<a href = "../../txt/gnu_lgpl.txt">the GNU LGPL license.</a>
</p>
<h3 align = "center">
Languages:
</h3>
<p>
<b>LCVT_DATASET</b> is available in
<a href = "../../cpp_src/lcvt_dataset/lcvt_dataset.html">a C++ version</a> and
<a href = "../../f_src/lcvt_dataset/lcvt_dataset.html">a FORTRAN90 version</a> and
<a href = "../../m_src/lcvt_dataset/lcvt_dataset.html">a MATLAB version.</a>
</p>
<h3 align = "center">
Related Data and Programs:
</h3>
<p>
<a href = "../../cpp_src/cvt/cvt.html">
CVT</a>,
a C++ library which
can compute a
CVT (Centroidal Voronoi Tessellation).
</p>
<p>
<a href = "../../cpp_src/cvt_dataset/cvt_dataset.html">
CVT_DATASET</a>,
a C++ program which
can create a CVT dataset (Centroidal Voronoi Tessellation).
</p>
<p>
<a href = "../../cpp_src/faure_dataset/faure_dataset.html">
FAURE_DATASET</a>,
a C++ program which
creates a Faure quasirandom dataset;
</p>
<p>
<a href = "../../cpp_src/grid_dataset/grid_dataset.html">
GRID_DATASET</a>,
a C++ program which
creates a grid sequence and writes it to a file.
</p>
<p>
<a href = "../../cpp_src/latin_center_dataset/latin_center_dataset.html">
LATIN_CENTER_DATASET</a>,
a C++ program which
creates a Latin Center Hypercube dataset;
</p>
<p>
<a href = "../../cpp_src/latin_edge_dataset/latin_edge_dataset.html">
LATIN_EDGE_DATASET</a>,
a C++ program which
creates a Latin Edge Hypercube dataset;
</p>
<p>
<a href = "../../cpp_src/latin_random_dataset/latin_random_dataset.html">
LATIN_RANDOM_DATASET</a>,
a C++ program which
creates a Latin Random Hypercube dataset;
</p>
<p>
<a href = "../../cpp_src/lcvt/lcvt.html">
LCVT</a>,
a C++ library which
is used by
<b>LCVT_DATASET</b>; a compiled copy of that library must be
available to build the program.
</p>
<p>
<a href = "../../datasets/lcvt/lcvt.html">
LCVT</a>,
a dataset directory which
contains a collection of sample
LCVT datasets created by <b>LCVT_DATASET</b>.
</p>
<p>
<a href = "../../cpp_src/niederreiter2_dataset/niederreiter2_dataset.html">
NIEDERREITER2_DATASET</a>,
a C++ program which
creates a Niederreiter quasirandom dataset with base 2;
</p>
<p>
<a href = "../../cpp_src/normal_dataset/normal_dataset.html">
NORMAL_DATASET</a>,
a C++ program which
generates a dataset of multivariate normal pseudorandom values and writes them to a file.
</p>
<p>
<a href = "../../cpp_src/sobol_dataset/sobol_dataset.html">
SOBOL_DATASET</a>,
a C++ program which
computes a Sobol quasirandom sequence and writes it to a file.
</p>
<p>
<a href = "../../cpp_src/table_latinize/table_latinize.html">
TABLE_LATINIZE</a>,
a C++ program which
can read a <b>TABLE file</b> of points and "latinize" the points,
that is, "gently" rearranging them so that they are regularly
spaced in every coordinate direction.
</p>
<p>
<a href = "../../cpp_src/uniform_dataset/uniform_dataset.html">
UNIFORM_DATASET</a>,
a C++ program which
generates a dataset of uniform pseudorandom values and writes them to a file.
</p>
<p>
<a href = "../../cpp_src/van_der_corput_dataset/van_der_corput_dataset.html">
VAN_DER_CORPUT_DATASET</a>,
a C++ program which
creates a van der Corput quasirandom sequence and writes it to a file.
</p>
<h3 align = "center">
Reference:
</h3>
<p>
<ol>
<li>
Franz Aurenhammer,<br>
Voronoi diagrams -
a study of a fundamental geometric data structure,<br>
ACM Computing Surveys,<br>
Volume 23, Number 3, September 1991, pages 345-405.
</li>
<li>
Franz Aurenhammer, Rolf Klein,<br>
Voronoi Diagrams,<br>
in Handbook of Computational Geometry,<br>
edited by J Sack, J Urrutia,<br>
Elsevier, 1999,<br>
LC: QA448.D38H36.
</li>
<li>
John Burkardt, Max Gunzburger, Janet Peterson, Rebecca Brannon,<br>
User Manual and Supporting Information for Library of Codes
for Centroidal Voronoi Placement and Associated Zeroth,
First, and Second Moment Determination,<br>
Sandia National Laboratories Technical Report SAND2002-0099,<br>
February 2002.
</li>
<li>
Qiang Du, Vance Faber, Max Gunzburger,<br>
Centroidal Voronoi Tessellations: Applications and Algorithms,<br>
SIAM Review,<br>
Volume 41, Number 4, December 1999, pages 637-676.
</li>
<li>
Vicente Romero, John Burkardt, Max Gunzburger, Janet Peterson, <br>
Initial Evaluation of Pure and "Latinized" Centroidal Voronoi
Tessellation for Non-Uniform Statistical Sampling,<br>
Sensitivity Analysis of Model Output (SAMO 2004) Conference,
Santa Fe, March 8-11, 2004.
</li>
<li>
Yuki Saka, Max Gunzburger, John Burkardt, <br>
Latinized, improved LHS, and CVT point sets in hypercubes, <br>
submitted to IEEE Transactions on Information Theory.
</li>
</ol>
</p>
<h3 align = "center">
Source Code:
</h3>
<p>
<ul>
<li>
<a href = "lcvt_dataset.cpp">lcvt_dataset.cpp</a>,
the source code.
</li>
<li>
<a href = "lcvt_dataset.sh">lcvt_dataset.sh</a>,
commands to compile, link and load the source code.
</li>
</ul>
</p>
<h3 align = "center">
Examples and Tests:
</h3>
<p>
<b>Example 1</b> is a dataset of N=85 points with spatial
dimension M=2, using UNIFORM initialization and sampling,
and 10,000 sample points:
<ul>
<li>
<a href = "lcvt01_input.txt">lcvt01_input.txt</a>,
input commands.
</li>
<li>
<a href = "lcvt01_output.txt">lcvt01_output.txt</a>,
printed output.
</li>
<li>
<a href = "lcvt01.txt">lcvt01.txt</a>,
the LCVT dataset.
</li>
<li>
<a href = "lcvt01.png">lcvt01.png</a>,
a <a href = "../../png/png.html">PNG</a> image of
the LCVT dataset,
created by
<a href =
"../../g_src/plot_points/plot_points.html">
PLOT_POINTS</a>.
</li>
</ul>
</p>
<p>
<b>Example 2</b> is a dataset of N=85 points with spatial
dimension M=2, using RANDOM initialization and sampling,
and 250,000 sample points, 10 CVT iterations
and 2 Latinization iterations:
<ul>
<li>
<a href = "lcvt02_input.txt">lcvt02_input.txt</a>,
input commands.
</li>
<li>
<a href = "lcvt02_output.txt">lcvt02_output.txt</a>,
printed output.
</li>
<li>
<a href = "lcvt02.txt">lcvt02.txt</a>,
the LCVT dataset.
</li>
<li>
<a href = "lcvt02.png">lcvt02.png</a>,
a <a href = "../../png/png.html">PNG</a> image of
the LCVT dataset,
created by
<a href =
"../../g_src/plot_points/plot_points.html">
PLOT_POINTS</a>.
</li>
</ul>
</p>
<p>
<b>Example 3</b> is a dataset of N=200 points with spatial
dimension M=7, using UNIFORM initialization and sampling,
and 20,000 sample points, 5 CVT iterations
and 2 Latinization iterations:
<ul>
<li>
<a href = "lcvt03_input.txt">lcvt03_input.txt</a>,
input commands.
</li>
<li>
<a href = "lcvt03_output.txt">lcvt03_output.txt</a>,
printed output.
</li>
<li>
<a href = "lcvt03.txt">lcvt03.txt</a>,
the LCVT dataset.
</li>
<li>
<a href = "lcvt03_page1.png">lcvt03_page1.png</a>,
"page 1" of
a <a href = "../../png/png.html">PNG</a> image of
pairs of coordinates of the LCVT dataset,
created by
<a href =
"../../f_src/table_top/table_top.html">
TABLE_TOP</a>.
</li>
<li>
<a href = "lcvt03_page2.png">lcvt03_page2.png</a>,
"page 2" of
a <a href = "../../png/png.html">PNG</a> image of
pairs of coordinates of the LCVT dataset,
created by
<a href =
"../../f_src/table_top/table_top.html">
TABLE_TOP</a>.
</li>
<li>
<a href = "lcvt03_page3.png">lcvt03_page3.png</a>,
"page 3" of
a <a href = "../../png/png.html">PNG</a> image of
pairs of coordinates of the LCVT dataset,
created by
<a href =
"../../f_src/table_top/table_top.html">
TABLE_TOP</a>.
</li>
<li>
<a href = "lcvt03_page4.png">lcvt03_page4.png</a>,
"page 4" of
a <a href = "../../png/png.html">PNG</a> image of
pairs of coordinates of the LCVT dataset,
created by
<a href =
"../../f_src/table_top/table_top.html">
TABLE_TOP</a>.
</li>
</ul>
</p>
<h3 align = "center">
List of Routines:
</h3>
<p>
<ul>
<li>
<b>MAIN</b> is the main program for LCVT_DATASET.
</li>
<li>
<b>CH_CAP</b> capitalizes a single character.
</li>
<li>
<b>CH_EQI</b> is true if two characters are equal, disregarding case.
</li>
<li>
<b>CH_TO_DIGIT</b> returns the integer value of a base 10 digit.
</li>
<li>
<b>CLUSTER_ENERGY</b> returns the energy of a dataset.
</li>
<li>
<b>CVT_ITERATION</b> takes one step of the CVT iteration.
</li>
<li>
<b>FIND_CLOSEST</b> finds the Voronoi cell generator closest to a point X.
</li>
<li>
<b>I4_MAX</b> returns the maximum of two I4's.
</li>
<li>
<b>I4_MIN</b> returns the smaller of two I4's.
</li>
<li>
<b>I4_TO_HALTON</b> computes an element of a Halton sequence.
</li>
<li>
<b>LCVT_WRITE</b> writes a Latinized CVT dataset to a file.
</li>
<li>
<b>PRIME</b> returns any of the first PRIME_MAX prime numbers.
</li>
<li>
<b>R8_EPSILON</b> returns the roundoff unit for R8 arithmetic.
</li>
<li>
<b>R8_UNIFORM_01</b> returns a unit pseudorandom R8.
</li>
<li>
<b>R8MAT_LATINIZE</b> "Latinizes" an R8MAT.
</li>
<li>
<b>R8TABLE_DATA_READ</b> reads the data from an R8TABLE file.
</li>
<li>
<b>R8VEC_SORT_HEAP_INDEX_A</b> does an indexed heap ascending sort of an R8VEC.
</li>
<li>
<b>REGION_SAMPLER</b> returns a sample point in the physical region.
</li>
<li>
<b>S_EQI</b> reports whether two strings are equal, ignoring case.
</li>
<li>
<b>S_LEN_TRIM</b> returns the length of a string to the last nonblank.
</li>
<li>
<b>S_TO_R8</b> reads an R8 from a string.
</li>
<li>
<b>S_TO_R8VEC</b> reads an R8VEC from a string.
</li>
<li>
<b>TIMESTAMP</b> prints the current YMDHMS date as a time stamp.
</li>
<li>
<b>TIMESTRING</b> returns the current YMDHMS date as a string.
</li>
<li>
<b>TUPLE_NEXT_FAST</b> computes the next element of a tuple space, "fast".
</li>
</ul>
</p>
<p>
You can go up one level to <a href = "../cpp_src.html">
the C++ source codes</a>.
</p>
<hr>
<i>
Last revised on 05 March 2007.
</i>
<!-- John Burkardt -->
</body>
<!-- Initial HTML skeleton created by HTMLINDEX. -->
</html>