forked from ivanku/linalg
-
Notifications
You must be signed in to change notification settings - Fork 1
/
README
443 lines (370 loc) · 12.9 KB
/
README
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
= Linalg - Ruby Linear Algebra Library
A Fortran-based linear algebra package.
=== Features
Major features:
* Cholesky decomposition
* LU decomposition
* QR decomposition
* Schur decomposition
* Singular value decomposition
* Eigenvalues and eigenvectors of a general matrix
* Minimization by least squares
* Linear equation solving
* Stand-alone LAPACK bindings: call any LAPACK routine from directly from ruby.
Minor features:
* Convenient iterators
* Condition numbers and condition number estimates
* Nullspace, rank, nullity
* Inverse
* Pseudo-inverse
* Determinant
* 2-norm, 1-norm, infinity-norm, Frobenius norm
=== Getting Started
Everything you need to know is in:
* Linalg::DMatrix
* Linalg::Iterators
and this README.
=== Tutorial
$ irb
irb(main):000:0> require 'linalg'
=> true
irb(main):000:0> include Linalg
=> Object
==== Construction
irb(main):000:0> DMatrix[[1,2,3], [4,5,6]]
=>
1.000000 2.000000 3.000000
4.000000 5.000000 6.000000
irb(main):000:0> DMatrix.rows [[1,2,3], [4,5,6]]
=>
1.000000 2.000000 3.000000
4.000000 5.000000 6.000000
irb(main):000:0> DMatrix.columns [[1,2,3], [4,5,6]]
=>
1.000000 4.000000
2.000000 5.000000
3.000000 6.000000
irb(main):000:0> a = DMatrix.new(3, 3) { |i, j| 10*i + j }
=>
0.000000 1.000000 2.000000
10.000000 11.000000 12.000000
20.000000 21.000000 22.000000
irb(main):000:0> DMatrix.new(3, 3, 99)
=>
99.000000 99.000000 99.000000
99.000000 99.000000 99.000000
99.000000 99.000000 99.000000
irb(main):000:0> DMatrix.diagonal [3,4,5]
=>
3.000000 0.000000 0.000000
0.000000 4.000000 0.000000
0.000000 0.000000 5.000000
irb(main):000:0> DMatrix.diagonal(4) { |i| i*i }
=>
0.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 4.000000 0.000000
0.000000 0.000000 0.000000 9.000000
irb(main):000:0> DMatrix.diagonal(4, 99)
=>
99.000000 0.000000 0.000000 0.000000
0.000000 99.000000 0.000000 0.000000
0.000000 0.000000 99.000000 0.000000
0.000000 0.000000 0.000000 99.000000
==== Indexing
Indexing is in (row, column) order. This is the convention for
Mathematics and Fortran. It is opposite from C convention.
irb(main):000:0> a
=>
0.000000 1.000000 2.000000
10.000000 11.000000 12.000000
20.000000 21.000000 22.000000
irb(main):000:0> a[1,0]
=> 10.0
irb(main):000:0> a[2,0]
=> 20.0
irb(main):000:0> a[0,1]
=> 1.0
irb(main):000:0> a[0,2]
=> 2.0
Index boundaries are strongly enforced
irb(main):000:0> a[-1,0]
IndexError: out of range
from (irb):27:in `[]'
from (irb):27
==== Enumerables
There are several abstract <tt>Enumerable</tt>s which you may obtain
from a matrix: columns, rows, elements, and diagonal elements.
irb(main):000:0> a
=>
0.000000 1.000000 2.000000
10.000000 11.000000 12.000000
20.000000 21.000000 22.000000
irb(main):000:0> a.columns.class
=> Linalg::Iterators::ColumnEnum
irb(main):000:0> cols = a.columns.map { |x| x }
=> [
0.000000
10.000000
20.000000
,
1.000000
11.000000
21.000000
,
2.000000
12.000000
22.000000
]
irb(main):000:0> rows = a.rows.map { |x| x }
=> [
0.000000 1.000000 2.000000
,
10.000000 11.000000 12.000000
,
20.000000 21.000000 22.000000
]
irb(main):000:0> a.elems.map { |x| x }
=> [0.0, 10.0, 20.0, 1.0, 11.0, 21.0, 2.0, 12.0, 22.0]
irb(main):003:0> a.elems.find_all { |x| x > 10 }
=> [20.0, 11.0, 21.0, 12.0, 22.0]
irb(main):008:0> a.diags.map { |x| x }
=> [0.0, 11.0, 22.0]
Another method of constructing a matrix is to join rows or columns,
irb(main):000:0> DMatrix.join_columns [cols[0], cols[2]]
=>
0.000000 2.000000
10.000000 12.000000
20.000000 22.000000
irb(main):000:0> DMatrix.join_rows [rows[0], rows[2]]
=>
0.000000 1.000000 2.000000
20.000000 21.000000 22.000000
==== Enumerable-like Iterators with Index Pairs
A matrix itself is not +Enumerable+, but a select number of
+Enumerable+-like methods are provided.
irb(main):000:0> a
=>
0.000000 1.000000 2.000000
10.000000 11.000000 12.000000
20.000000 21.000000 22.000000
irb(main):000:0> a.each_with_index { |e, i, j| puts "row #{i} column #{j} : #{e}" } ; nil
row 0 column 0 : 0.0
row 1 column 0 : 10.0
row 2 column 0 : 20.0
row 0 column 1 : 1.0
row 1 column 1 : 11.0
row 2 column 1 : 21.0
row 0 column 2 : 2.0
row 1 column 2 : 12.0
row 2 column 2 : 22.0
=> nil
irb(main):000:0> a.map_with_index { |e, i, j| e*i*j }
=>
0.000000 0.000000 0.000000
0.000000 11.000000 24.000000
0.000000 42.000000 88.000000
irb(main):000:0> a.each_upper_with_index { |e, i, j| puts "a[#{i}, #{j}] : #{e}" } ; nil
a[0, 1] : 1.0
a[0, 2] : 2.0
a[1, 2] : 12.0
=> nil
irb(main):000:0> a.each_lower_with_index { |e, i, j| puts "a[#{i}, #{j}] : #{e}" } ; nil
a[1, 0] : 10.0
a[2, 0] : 20.0
a[2, 1] : 21.0
=> nil
==== Epsilon Comparison
For good and bad, a default epsilon of <tt>1e-8</tt> is provided for
comparison, nullspace identification, and symmetric testing.
You can change +default_epsilon+ class-wide or on a per-object basis,
or simply pass an explicit epsilon to any of these methods.
irb(main):000:0> a = DMatrix.rand(3, 3)
=>
0.824730 0.305527 0.044433
-0.582865 -0.351364 -0.752941
0.103417 -0.254290 0.216312
irb(main):000:0> b = a.map { |e| e + 0.000001 }
=>
0.824731 0.305528 0.044434
-0.582864 -0.351363 -0.752940
0.103418 -0.254289 0.216313
irb(main):000:0> a.within(1e-4, b)
=> true
irb(main):000:0> a.class.default_epsilon
=> 1.0e-08
irb(main):000:0> a =~ b
=> false
irb(main):000:0> a.singleton_class.default_epsilon
=> nil
irb(main):000:0> a.singleton_class.default_epsilon = 0.0001
=> 0.0001
irb(main):000:0> a =~ b
=> true
irb(main):000:0> b =~ a
=> false
<tt>singleton_class.epsilon</tt> has first preference over
<tt>class.epsilon</tt>.
==== Singular Value Decomposition
irb(main):000:0> a = DMatrix.rand(4, 7) ;
irb(main):000:0* u, s, vt = a.singular_value_decomposition
=> [
-0.747003 0.304315 -0.144972 -0.573029
-0.435034 -0.814506 0.381951 0.037926
0.207010 -0.490811 -0.777727 -0.333753
-0.458125 0.055467 -0.477741 0.747535
,
2.186983 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
0.000000 1.719562 0.000000 0.000000 0.000000 0.000000 0.000000
0.000000 0.000000 1.474243 0.000000 0.000000 0.000000 0.000000
0.000000 0.000000 0.000000 0.676138 0.000000 0.000000 0.000000
,
0.276463 -0.345917 0.573929 -0.416910 -0.139648 0.526564 0.062694
-0.838456 -0.430716 0.199266 0.170144 0.087954 0.077829 0.170372
-0.022382 -0.403705 -0.054483 -0.077530 -0.042506 -0.160232 -0.894461
-0.079171 0.476099 0.396614 0.253131 0.573392 0.315335 -0.342737
0.034285 -0.212892 -0.265934 -0.537325 0.749204 -0.111692 0.142407
-0.377150 0.344083 -0.443538 -0.422834 -0.235075 0.530130 -0.165989
-0.265280 0.376113 0.450755 -0.509553 -0.160031 -0.545940 -0.041002
]
irb(main):000:0> u*s*vt
=>
-0.854950 0.241548 -0.975366 0.688628 0.061092 -0.907441 0.310691
0.896671 0.717254 -0.845643 0.121186 0.000445 -0.692125 -0.810721
0.876331 0.562343 0.064623 -0.300574 -0.218113 0.285261 0.987489
-0.381214 0.830467 -0.317184 0.616481 0.468055 -0.247913 0.410178
irb(main):000:0> a
=>
-0.854950 0.241548 -0.975366 0.688628 0.061092 -0.907441 0.310691
0.896671 0.717254 -0.845643 0.121186 0.000445 -0.692125 -0.810721
0.876331 0.562343 0.064623 -0.300574 -0.218113 0.285261 0.987489
-0.381214 0.830467 -0.317184 0.616481 0.468055 -0.247913 0.410178
irb(main):000:0> u*u.t
=>
1.000000 0.000000 -0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
-0.000000 0.000000 1.000000 0.000000
0.000000 0.000000 0.000000 1.000000
irb(main):000:0> vt.t*vt
=>
1.000000 0.000000 0.000000 0.000000 0.000000 0.000000 -0.000000
0.000000 1.000000 -0.000000 -0.000000 0.000000 -0.000000 0.000000
0.000000 -0.000000 1.000000 -0.000000 -0.000000 0.000000 0.000000
0.000000 -0.000000 -0.000000 1.000000 -0.000000 -0.000000 0.000000
0.000000 0.000000 -0.000000 -0.000000 1.000000 -0.000000 0.000000
0.000000 -0.000000 0.000000 -0.000000 -0.000000 1.000000 -0.000000
-0.000000 0.000000 0.000000 0.000000 0.000000 -0.000000 1.000000
==== Eigenvectors and Eigenvalues
irb(main):000:0> a = DMatrix.rand(5, 5)
=>
-0.319566 0.633985 0.335298 -0.150403 0.758559
-0.633389 0.444269 0.375873 0.521107 0.247966
0.757654 0.504831 0.160970 -0.241885 -0.949746
-0.174517 0.351239 -0.600079 -0.533921 0.851118
-0.736717 0.006612 -0.941311 -0.417801 0.555841
irb(main):000:0> eigs, re, im = a.eigensystem
=> [
-0.232169 -0.540550 -0.215201 -0.216959 0.036677
-0.449048 -0.031420 -0.114366 -0.228818 0.062841
0.682632 0.283575 0.019432 0.540907 0.073854
0.364843 -0.597905 0.000000 -0.592036 0.000000
0.381257 -0.207285 -0.407649 -0.490737 -0.076896
,
-1.088525
-0.093563
-0.093563
0.791621
0.791621
,
0.000000
0.604163
-0.604163
0.158934
-0.158934
]
irb(main):000:0> a*eigs.column(0)
=>
0.252722
0.488799
-0.743062
-0.397141
-0.415008
irb(main):000:0> re[0]*eigs.column(0)
=>
0.252722
0.488799
-0.743062
-0.397141
-0.415008
==== QR factorization
irb(main):000:0> a = DMatrix.rand(4, 7) ;
irb(main):000:0* q, r = a.qr
=> [
-0.593983 0.263367 -0.572195 -0.500414
-0.486715 -0.780258 0.331047 -0.211458
-0.184137 -0.328403 -0.586889 0.716803
-0.613503 0.462588 0.467507 0.437109
,
-1.180464 -0.295897 -0.478813 -0.300434 0.360094 0.738799 -0.571687
0.000000 0.873800 -0.167873 0.612545 0.462280 0.097813 -0.733398
0.000000 0.000000 1.659791 -0.439355 -0.234388 -0.063566 0.149709
0.000000 0.000000 0.000000 -0.047795 0.805122 -0.158261 0.754104
]
irb(main):000:0> q*r
=>
0.701176 0.405888 -0.709530 0.615091 -0.360919 -0.297505 -0.316607
0.574550 -0.537772 0.913498 -0.467057 -0.783803 -0.423482 0.740588
0.217367 -0.232473 -0.830816 0.077752 0.496553 -0.244298 0.798800
0.724218 0.585743 0.992061 0.241379 0.235274 -0.506903 0.411086
irb(main):000:0> a
=>
0.701176 0.405888 -0.709530 0.615091 -0.360919 -0.297505 -0.316607
0.574550 -0.537772 0.913498 -0.467057 -0.783803 -0.423482 0.740588
0.217367 -0.232473 -0.830816 0.077752 0.496553 -0.244298 0.798800
0.724218 0.585743 0.992061 0.241379 0.235274 -0.506903 0.411086
irb(main):000:0> q.t*q
=>
1.000000 0.000000 -0.000000 -0.000000
0.000000 1.000000 0.000000 -0.000000
-0.000000 0.000000 1.000000 0.000000
-0.000000 -0.000000 0.000000 1.000000
See the Linalg::DMatrix documentation for more info. The various
tests in test/ are also instructive.
=== Download
* http://rubyforge.org/frs/?group_id=273
=== Repository
* http://github.com/quix/linalg
=== Notes
There are four matrix types: +SMatrix+, +DMatrix+, +CMatrix+, and
+ZMatrix+ -- single precision, double precision, single precision
complex, and double precision complex, respectively. They are all
available with basic functionality, however the more complex routines
you see here currently lie only in +DMatrix+.
If you have used +narray+, note that +linalg+ uses the mathematical
definition of <em>rank</em>, which is equal to the number of columns
only in the case of a nonsingular square matrix.
=== Details
Author:: James M. Lawrence <[email protected]>
Requires:: Ruby 1.8.1 or later
License:: Copyright (c) 2004-2008 James M. Lawrence.
Released under the MIT license.
=== License
Copyright (c) 2004-2008 James M. Lawrence
If +linalg+ begins to smoke, get away immediately. Seek shelter and
cover head.
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.