nimble/UserManual/cha-writing-models.html at devel · nimble-dev/nimble · GitHub

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
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html lang="" xml:lang="">
<head>

  <meta charset="utf-8" />
  <meta http-equiv="X-UA-Compatible" content="IE=edge" />
  <title>Chapter 5 Writing models in NIMBLE’s dialect of BUGS | NimbleUserManual.knit</title>
  <meta name="description" content="This is the NIMBLE User Manual." />
  <meta name="generator" content="bookdown 0.37 and GitBook 2.6.7" />

  <meta property="og:title" content="Chapter 5 Writing models in NIMBLE’s dialect of BUGS | NimbleUserManual.knit" />
  <meta property="og:type" content="book" />
  <meta property="og:image" content="/nimble-icon.png" />
  <meta property="og:description" content="This is the NIMBLE User Manual." />
  <meta name="github-repo" content="nimble-dev/nimble" />

  <meta name="twitter:card" content="summary" />
  <meta name="twitter:title" content="Chapter 5 Writing models in NIMBLE’s dialect of BUGS | NimbleUserManual.knit" />

  <meta name="twitter:description" content="This is the NIMBLE User Manual." />
  <meta name="twitter:image" content="/nimble-icon.png" />


  <meta name="viewport" content="width=device-width, initial-scale=1" />
  <meta name="apple-mobile-web-app-capable" content="yes" />
  <meta name="apple-mobile-web-app-status-bar-style" content="black" />


<link rel="prev" href="cha-installing-nimble.html"/>
<link rel="next" href="cha-building-models.html"/>
<script src="libs/jquery-3.6.0/jquery-3.6.0.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/fuse.js@6.4.6/dist/fuse.min.js"></script>
<link href="libs/gitbook-2.6.7/css/style.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-table.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-bookdown.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-highlight.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-search.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-fontsettings.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-clipboard.css" rel="stylesheet" />


<link href="libs/anchor-sections-1.1.0/anchor-sections.css" rel="stylesheet" />
<link href="libs/anchor-sections-1.1.0/anchor-sections-hash.css" rel="stylesheet" />
<script src="libs/anchor-sections-1.1.0/anchor-sections.js"></script>


<style type="text/css">
pre > code.sourceCode { white-space: pre; position: relative; }
pre > code.sourceCode > span { line-height: 1.25; }
pre > code.sourceCode > span:empty { height: 1.2em; }
.sourceCode { overflow: visible; }
code.sourceCode > span { color: inherit; text-decoration: inherit; }
pre.sourceCode { margin: 0; }
@media screen {
div.sourceCode { overflow: auto; }
}
@media print {
pre > code.sourceCode { white-space: pre-wrap; }
pre > code.sourceCode > span { text-indent: -5em; padding-left: 5em; }
}
pre.numberSource code
  { counter-reset: source-line 0; }
pre.numberSource code > span
  { position: relative; left: -4em; counter-increment: source-line; }
pre.numberSource code > span > a:first-child::before
  { content: counter(source-line);
    position: relative; left: -1em; text-align: right; vertical-align: baseline;
    border: none; display: inline-block;
    -webkit-touch-callout: none; -webkit-user-select: none;
    -khtml-user-select: none; -moz-user-select: none;
    -ms-user-select: none; user-select: none;
    padding: 0 4px; width: 4em;
    color: #aaaaaa;
  }
pre.numberSource { margin-left: 3em; border-left: 1px solid #aaaaaa;  padding-left: 4px; }
div.sourceCode
  {   }
@media screen {
pre > code.sourceCode > span > a:first-child::before { text-decoration: underline; }
}
code span.al { color: #ff0000; font-weight: bold; } /* Alert */
code span.an { color: #60a0b0; font-weight: bold; font-style: italic; } /* Annotation */
code span.at { color: #7d9029; } /* Attribute */
code span.bn { color: #40a070; } /* BaseN */
code span.bu { color: #008000; } /* BuiltIn */
code span.cf { color: #007020; font-weight: bold; } /* ControlFlow */
code span.ch { color: #4070a0; } /* Char */
code span.cn { color: #880000; } /* Constant */
code span.co { color: #60a0b0; font-style: italic; } /* Comment */
code span.cv { color: #60a0b0; font-weight: bold; font-style: italic; } /* CommentVar */
code span.do { color: #ba2121; font-style: italic; } /* Documentation */
code span.dt { color: #902000; } /* DataType */
code span.dv { color: #40a070; } /* DecVal */
code span.er { color: #ff0000; font-weight: bold; } /* Error */
code span.ex { } /* Extension */
code span.fl { color: #40a070; } /* Float */
code span.fu { color: #06287e; } /* Function */
code span.im { color: #008000; font-weight: bold; } /* Import */
code span.in { color: #60a0b0; font-weight: bold; font-style: italic; } /* Information */
code span.kw { color: #007020; font-weight: bold; } /* Keyword */
code span.op { color: #666666; } /* Operator */
code span.ot { color: #007020; } /* Other */
code span.pp { color: #bc7a00; } /* Preprocessor */
code span.sc { color: #4070a0; } /* SpecialChar */
code span.ss { color: #bb6688; } /* SpecialString */
code span.st { color: #4070a0; } /* String */
code span.va { color: #19177c; } /* Variable */
code span.vs { color: #4070a0; } /* VerbatimString */
code span.wa { color: #60a0b0; font-weight: bold; font-style: italic; } /* Warning */
</style>

<style type="text/css">

  div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
</style>
<style type="text/css">
/* Used with Pandoc 2.11+ new --citeproc when CSL is used */
div.csl-bib-body { }
div.csl-entry {
  clear: both;
}
.hanging div.csl-entry {
  margin-left:2em;
  text-indent:-2em;
}
div.csl-left-margin {
  min-width:2em;
  float:left;
}
div.csl-right-inline {
  margin-left:2em;
  padding-left:1em;
}
div.csl-indent {
  margin-left: 2em;
}
</style>

<link rel="stylesheet" href="style.css" type="text/css" />
</head>

<body>


  <div class="book without-animation with-summary font-size-2 font-family-1" data-basepath=".">

    <div class="book-summary">
      <nav role="navigation">

<ul class="summary">
<img src="./nimble-icon.png"
     width=100>
<li><a href="./cha-welcome-nimble.html">NIMBLE User Manual, Version 1.4.2</a></li>
<li><a href="https://github.com/nimble-dev/nimble">NIMBLE Development Team</a></li>
<li><a href="https://R-nimble.org">https://R-nimble.org</a></li>

<li class="divider"></li>
<li class="part"><span><b>I Introduction</b></span></li>
<li class="chapter" data-level="1" data-path="cha-welcome-nimble.html"><a href="cha-welcome-nimble.html"><i class="fa fa-check"></i><b>1</b> Welcome to NIMBLE</a>
<ul>
<li class="chapter" data-level="1.1" data-path="cha-welcome-nimble.html"><a href="cha-welcome-nimble.html#sec:what-is-nimble"><i class="fa fa-check"></i><b>1.1</b> What does NIMBLE do?</a></li>
<li class="chapter" data-level="1.2" data-path="cha-welcome-nimble.html"><a href="cha-welcome-nimble.html#how-to-use-this-manual"><i class="fa fa-check"></i><b>1.2</b> How to use this manual</a></li>
</ul></li>
<li class="chapter" data-level="2" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html"><i class="fa fa-check"></i><b>2</b> Lightning introduction</a>
<ul>
<li class="chapter" data-level="2.1" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:brief-example"><i class="fa fa-check"></i><b>2.1</b> A brief example</a></li>
<li class="chapter" data-level="2.2" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:creating-model"><i class="fa fa-check"></i><b>2.2</b> Creating a model</a></li>
<li class="chapter" data-level="2.3" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:compiling-model"><i class="fa fa-check"></i><b>2.3</b> Compiling the model</a></li>
<li class="chapter" data-level="2.4" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:intro-runMCMC"><i class="fa fa-check"></i><b>2.4</b> One-line invocation of MCMC</a></li>
<li class="chapter" data-level="2.5" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:creating-mcmc"><i class="fa fa-check"></i><b>2.5</b> Creating, compiling and running a basic MCMC configuration</a></li>
<li class="chapter" data-level="2.6" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:customizing-mcmc"><i class="fa fa-check"></i><b>2.6</b> Customizing the MCMC</a></li>
<li class="chapter" data-level="2.7" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:running-mcem"><i class="fa fa-check"></i><b>2.7</b> Running MCEM</a></li>
<li class="chapter" data-level="2.8" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:creating-your-own"><i class="fa fa-check"></i><b>2.8</b> Creating your own functions</a></li>
</ul></li>
<li class="chapter" data-level="3" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html"><i class="fa fa-check"></i><b>3</b> More introduction</a>
<ul>
<li class="chapter" data-level="3.1" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html#nimble-adopts-and-extends-the-bugs-language-for-specifying-models"><i class="fa fa-check"></i><b>3.1</b> NIMBLE adopts and extends the BUGS language for specifying models</a></li>
<li class="chapter" data-level="3.2" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html#sec:nimble-lang-writ"><i class="fa fa-check"></i><b>3.2</b> nimbleFunctions for writing algorithms</a></li>
<li class="chapter" data-level="3.3" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html#sec:nimble-algor-libr"><i class="fa fa-check"></i><b>3.3</b> The NIMBLE algorithm library</a></li>
</ul></li>
<li class="chapter" data-level="4" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html"><i class="fa fa-check"></i><b>4</b> Installing NIMBLE</a>
<ul>
<li class="chapter" data-level="4.1" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#sec:requ-run-nimble"><i class="fa fa-check"></i><b>4.1</b> Requirements to run NIMBLE</a></li>
<li class="chapter" data-level="4.2" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#sec:compiler"><i class="fa fa-check"></i><b>4.2</b> Installing a C++ compiler for NIMBLE to use</a>
<ul>
<li class="chapter" data-level="4.2.1" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#macos"><i class="fa fa-check"></i><b>4.2.1</b> MacOS</a></li>
<li class="chapter" data-level="4.2.2" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#linux"><i class="fa fa-check"></i><b>4.2.2</b> Linux</a></li>
<li class="chapter" data-level="4.2.3" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#windows"><i class="fa fa-check"></i><b>4.2.3</b> Windows</a></li>
</ul></li>
<li class="chapter" data-level="4.3" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#installing-the-nimble-package"><i class="fa fa-check"></i><b>4.3</b> Installing the NIMBLE package</a></li>
<li class="chapter" data-level="4.4" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#troubleshooting-installation-problems"><i class="fa fa-check"></i><b>4.4</b> Troubleshooting installation problems</a></li>
<li class="chapter" data-level="4.5" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#customizing-your-installation"><i class="fa fa-check"></i><b>4.5</b> Customizing your installation</a>
<ul>
<li class="chapter" data-level="4.5.1" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#using-your-own-copy-of-eigen"><i class="fa fa-check"></i><b>4.5.1</b> Using your own copy of Eigen</a></li>
<li class="chapter" data-level="4.5.2" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#using-libnimble"><i class="fa fa-check"></i><b>4.5.2</b> Using libnimble</a></li>
<li class="chapter" data-level="4.5.3" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#sec:blas"><i class="fa fa-check"></i><b>4.5.3</b> BLAS and LAPACK</a></li>
<li class="chapter" data-level="4.5.4" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#customizing-compilation-of-the-nimble-generated-c"><i class="fa fa-check"></i><b>4.5.4</b> Customizing compilation of the NIMBLE-generated C++</a></li>
</ul></li>
</ul></li>
<li class="part"><span><b>II Models in NIMBLE</b></span></li>
<li class="chapter" data-level="5" data-path="cha-writing-models.html"><a href="cha-writing-models.html"><i class="fa fa-check"></i><b>5</b> Writing models in NIMBLE’s dialect of BUGS</a>
<ul>
<li class="chapter" data-level="5.1" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:supp-feat-bugs"><i class="fa fa-check"></i><b>5.1</b> Comparison to BUGS dialects supported by WinBUGS, OpenBUGS and JAGS</a>
<ul>
<li class="chapter" data-level="5.1.1" data-path="cha-writing-models.html"><a href="cha-writing-models.html#supported-features-of-bugs-and-jags"><i class="fa fa-check"></i><b>5.1.1</b> Supported features of BUGS and JAGS</a></li>
<li class="chapter" data-level="5.1.2" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:extensions-bugs"><i class="fa fa-check"></i><b>5.1.2</b> NIMBLE’s Extensions to BUGS and JAGS</a></li>
<li class="chapter" data-level="5.1.3" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:not-yet-supported"><i class="fa fa-check"></i><b>5.1.3</b> Not-supported features of BUGS and JAGS</a></li>
</ul></li>
<li class="chapter" data-level="5.2" data-path="cha-writing-models.html"><a href="cha-writing-models.html#writing-models"><i class="fa fa-check"></i><b>5.2</b> Writing models</a>
<ul>
<li class="chapter" data-level="5.2.1" data-path="cha-writing-models.html"><a href="cha-writing-models.html#declaring-stochastic-and-deterministic-nodes"><i class="fa fa-check"></i><b>5.2.1</b> Declaring stochastic and deterministic nodes</a></li>
<li class="chapter" data-level="5.2.2" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:more-kinds-bugs"><i class="fa fa-check"></i><b>5.2.2</b> More kinds of BUGS declarations</a></li>
<li class="chapter" data-level="5.2.3" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:vectorized-versus-scalar-declarations"><i class="fa fa-check"></i><b>5.2.3</b> Vectorized versus scalar declarations</a></li>
<li class="chapter" data-level="5.2.4" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:dists-and-functions"><i class="fa fa-check"></i><b>5.2.4</b> Available distributions</a></li>
<li class="chapter" data-level="5.2.5" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:BUGS-lang-fxns"><i class="fa fa-check"></i><b>5.2.5</b> Available BUGS language functions</a></li>
<li class="chapter" data-level="5.2.6" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:BUGS-link"><i class="fa fa-check"></i><b>5.2.6</b> Available link functions</a></li>
<li class="chapter" data-level="5.2.7" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:trunc"><i class="fa fa-check"></i><b>5.2.7</b> Truncation, censoring, and constraints</a></li>
<li class="chapter" data-level="5.2.8" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:macros"><i class="fa fa-check"></i><b>5.2.8</b> Model macros</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="6" data-path="cha-building-models.html"><a href="cha-building-models.html"><i class="fa fa-check"></i><b>6</b> Building and using models</a>
<ul>
<li class="chapter" data-level="6.1" data-path="cha-building-models.html"><a href="cha-building-models.html#creating-model-objects"><i class="fa fa-check"></i><b>6.1</b> Creating model objects</a>
<ul>
<li class="chapter" data-level="6.1.1" data-path="cha-building-models.html"><a href="cha-building-models.html#using-nimblemodel-to-create-a-model"><i class="fa fa-check"></i><b>6.1.1</b> Using <em>nimbleModel</em> to create a model</a></li>
<li class="chapter" data-level="6.1.2" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:readBUGSmodel"><i class="fa fa-check"></i><b>6.1.2</b> Creating a model from standard BUGS and JAGS input files</a></li>
<li class="chapter" data-level="6.1.3" data-path="cha-building-models.html"><a href="cha-building-models.html#sub:multiple-instances"><i class="fa fa-check"></i><b>6.1.3</b> Making multiple instances from the same model definition</a></li>
</ul></li>
<li class="chapter" data-level="6.2" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:nodes-and-variables"><i class="fa fa-check"></i><b>6.2</b> NIMBLE models are objects you can query and manipulate</a>
<ul>
<li class="chapter" data-level="6.2.1" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:what-are-nodes-and-variables"><i class="fa fa-check"></i><b>6.2.1</b> What are variables and nodes?</a></li>
<li class="chapter" data-level="6.2.2" data-path="cha-building-models.html"><a href="cha-building-models.html#determining-the-nodes-and-variables-in-a-model"><i class="fa fa-check"></i><b>6.2.2</b> Determining the nodes and variables in a model</a></li>
<li class="chapter" data-level="6.2.3" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:accessing-nodes"><i class="fa fa-check"></i><b>6.2.3</b> Accessing nodes</a></li>
<li class="chapter" data-level="6.2.4" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:how-nodes-are"><i class="fa fa-check"></i><b>6.2.4</b> How nodes are named</a></li>
<li class="chapter" data-level="6.2.5" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:why-use-node"><i class="fa fa-check"></i><b>6.2.5</b> Why use node names?</a></li>
<li class="chapter" data-level="6.2.6" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:cdisdata"><i class="fa fa-check"></i><b>6.2.6</b> Checking if a node holds data</a></li>
</ul></li>
<li class="chapter" data-level="6.3" data-path="cha-building-models.html"><a href="cha-building-models.html#using-models-in-parallel"><i class="fa fa-check"></i><b>6.3</b> Using models in parallel</a></li>
</ul></li>
<li class="part"><span><b>III Algorithms in NIMBLE</b></span></li>
<li class="chapter" data-level="7" data-path="cha-mcmc.html"><a href="cha-mcmc.html"><i class="fa fa-check"></i><b>7</b> MCMC</a>
<ul>
<li class="chapter" data-level="7.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:nimbleMCMC"><i class="fa fa-check"></i><b>7.1</b> One-line invocation of MCMC: <em>nimbleMCMC</em></a></li>
<li class="chapter" data-level="7.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:mcmc-configuration"><i class="fa fa-check"></i><b>7.2</b> The MCMC configuration</a>
<ul>
<li class="chapter" data-level="7.2.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:default-mcmc-conf"><i class="fa fa-check"></i><b>7.2.1</b> Default MCMC configuration</a></li>
<li class="chapter" data-level="7.2.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:customizing-mcmc-conf"><i class="fa fa-check"></i><b>7.2.2</b> Customizing the MCMC configuration</a></li>
<li class="chapter" data-level="7.2.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:derived-quantities"><i class="fa fa-check"></i><b>7.2.3</b> Setting up derived quantities for additional quantities of interest</a></li>
</ul></li>
<li class="chapter" data-level="7.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:build-compile-mcmc"><i class="fa fa-check"></i><b>7.3</b> Building and compiling the MCMC</a></li>
<li class="chapter" data-level="7.4" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:initMCMC"><i class="fa fa-check"></i><b>7.4</b> Initializing MCMC</a></li>
<li class="chapter" data-level="7.5" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:runMCMC"><i class="fa fa-check"></i><b>7.5</b> User-friendly execution of MCMC algorithms: <em>runMCMC</em></a></li>
<li class="chapter" data-level="7.6" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:executing-the-mcmc-algorithm"><i class="fa fa-check"></i><b>7.6</b> Running the MCMC</a>
<ul>
<li class="chapter" data-level="7.6.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:mcmc-rerun"><i class="fa fa-check"></i><b>7.6.1</b> Rerunning versus restarting an MCMC</a></li>
<li class="chapter" data-level="7.6.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:sampler-time"><i class="fa fa-check"></i><b>7.6.2</b> Measuring sampler computation times: <em>getTimes</em></a></li>
<li class="chapter" data-level="7.6.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#assessing-the-adaption-process-of-rw-and-rw_block-samplers"><i class="fa fa-check"></i><b>7.6.3</b> Assessing the adaption process of <em>RW</em> and <em>RW_block</em> samplers</a></li>
</ul></li>
<li class="chapter" data-level="7.7" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:extracting-samples"><i class="fa fa-check"></i><b>7.7</b> Extracting MCMC samples</a></li>
<li class="chapter" data-level="7.8" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:WAIC"><i class="fa fa-check"></i><b>7.8</b> Calculating WAIC</a></li>
<li class="chapter" data-level="7.9" data-path="cha-mcmc.html"><a href="cha-mcmc.html#k-fold-cross-validation"><i class="fa fa-check"></i><b>7.9</b> k-fold cross-validation</a></li>
<li class="chapter" data-level="7.10" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:rjmcmc"><i class="fa fa-check"></i><b>7.10</b> Variable selection using Reversible Jump MCMC</a>
<ul>
<li class="chapter" data-level="7.10.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:rjmcmc-indicator"><i class="fa fa-check"></i><b>7.10.1</b> Using indicator variables</a></li>
<li class="chapter" data-level="7.10.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:rjmcmc-no-indicator"><i class="fa fa-check"></i><b>7.10.2</b> Without indicator variables</a></li>
</ul></li>
<li class="chapter" data-level="7.11" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:samplers-provided"><i class="fa fa-check"></i><b>7.11</b> Samplers provided with NIMBLE</a>
<ul>
<li class="chapter" data-level="7.11.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#conjugate-gibbs-samplers"><i class="fa fa-check"></i><b>7.11.1</b> Conjugate (‘Gibbs’) samplers</a></li>
<li class="chapter" data-level="7.11.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#subsec:HMC"><i class="fa fa-check"></i><b>7.11.2</b> Hamiltonian Monte Carlo (HMC)</a></li>
<li class="chapter" data-level="7.11.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#particle-filter-samplers"><i class="fa fa-check"></i><b>7.11.3</b> Particle filter samplers</a></li>
<li class="chapter" data-level="7.11.4" data-path="cha-mcmc.html"><a href="cha-mcmc.html#customized-log-likelihood-evaluations-rw_llfunction-sampler"><i class="fa fa-check"></i><b>7.11.4</b> Customized log-likelihood evaluations: <em>RW_llFunction sampler</em></a></li>
</ul></li>
<li class="chapter" data-level="7.12" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:mcmc-example-litters"><i class="fa fa-check"></i><b>7.12</b> Detailed MCMC example: <em>litters</em></a></li>
<li class="chapter" data-level="7.13" data-path="cha-mcmc.html"><a href="cha-mcmc.html#mcmc-suite-compare-mcmcs"><i class="fa fa-check"></i><b>7.13</b> Comparing different MCMCs with <em>compareMCMCs</em></a></li>
<li class="chapter" data-level="7.14" data-path="cha-mcmc.html"><a href="cha-mcmc.html#running-mcmc-chains-in-parallel"><i class="fa fa-check"></i><b>7.14</b> Running MCMC chains in parallel</a></li>
</ul></li>
<li class="chapter" data-level="8" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html"><i class="fa fa-check"></i><b>8</b> Particle Filters, PMCMC, and MCEM</a>
<ul>
<li class="chapter" data-level="8.1" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#particle-filters-sequential-monte-carlo-and-iterated-filtering"><i class="fa fa-check"></i><b>8.1</b> Particle filters / sequential Monte Carlo and iterated filtering</a>
<ul>
<li class="chapter" data-level="8.1.1" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#filtering-algorithms"><i class="fa fa-check"></i><b>8.1.1</b> Filtering algorithms</a></li>
<li class="chapter" data-level="8.1.2" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#sec:particle-mcmc"><i class="fa fa-check"></i><b>8.1.2</b> Particle MCMC (PMCMC)</a></li>
</ul></li>
<li class="chapter" data-level="8.2" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#monte-carlo-expectation-maximization-mcem"><i class="fa fa-check"></i><b>8.2</b> Monte Carlo Expectation Maximization (MCEM)</a>
<ul>
<li class="chapter" data-level="8.2.1" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#sec:estimate-mcem-cov"><i class="fa fa-check"></i><b>8.2.1</b> Estimating the asymptotic covariance From MCEM</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="9" data-path="cha-laplace.html"><a href="cha-laplace.html"><i class="fa fa-check"></i><b>9</b> Laplace, AGHQ, and nested approximations</a>
<ul>
<li class="chapter" data-level="9.1" data-path="cha-laplace.html"><a href="cha-laplace.html#sec:AD-laplace"><i class="fa fa-check"></i><b>9.1</b> Laplace approximation and adaptive Gauss-Hermite quadrature (AGHQ)</a>
<ul>
<li class="chapter" data-level="9.1.1" data-path="cha-laplace.html"><a href="cha-laplace.html#glmm-example"><i class="fa fa-check"></i><b>9.1.1</b> GLMM example</a></li>
<li class="chapter" data-level="9.1.2" data-path="cha-laplace.html"><a href="cha-laplace.html#using-laplace-approximation"><i class="fa fa-check"></i><b>9.1.2</b> Using Laplace approximation</a></li>
<li class="chapter" data-level="9.1.3" data-path="cha-laplace.html"><a href="cha-laplace.html#using-the-laplace-approximation-methods-directly"><i class="fa fa-check"></i><b>9.1.3</b> Using the Laplace approximation methods directly</a></li>
<li class="chapter" data-level="9.1.4" data-path="cha-laplace.html"><a href="cha-laplace.html#changing-the-optimization-methods"><i class="fa fa-check"></i><b>9.1.4</b> Changing the optimization methods</a></li>
</ul></li>
<li class="chapter" data-level="9.2" data-path="cha-laplace.html"><a href="cha-laplace.html#sec:nested-approx"><i class="fa fa-check"></i><b>9.2</b> Nested approximation (INLA-like) methods</a>
<ul>
<li class="chapter" data-level="9.2.1" data-path="cha-laplace.html"><a href="cha-laplace.html#overview-of-the-methodology"><i class="fa fa-check"></i><b>9.2.1</b> Overview of the methodology</a></li>
<li class="chapter" data-level="9.2.2" data-path="cha-laplace.html"><a href="cha-laplace.html#using-nimbles-nested-approximation"><i class="fa fa-check"></i><b>9.2.2</b> Using NIMBLE’s nested approximation</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="10" data-path="cha-spatial.html"><a href="cha-spatial.html"><i class="fa fa-check"></i><b>10</b> Spatial models</a>
<ul>
<li class="chapter" data-level="10.1" data-path="cha-spatial.html"><a href="cha-spatial.html#intrinsic-gaussian-car-model-dcar_normal"><i class="fa fa-check"></i><b>10.1</b> Intrinsic Gaussian CAR model: <em>dcar_normal</em></a>
<ul>
<li class="chapter" data-level="10.1.1" data-path="cha-spatial.html"><a href="cha-spatial.html#specification-and-density"><i class="fa fa-check"></i><b>10.1.1</b> Specification and density</a></li>
<li class="chapter" data-level="10.1.2" data-path="cha-spatial.html"><a href="cha-spatial.html#example"><i class="fa fa-check"></i><b>10.1.2</b> Example</a></li>
</ul></li>
<li class="chapter" data-level="10.2" data-path="cha-spatial.html"><a href="cha-spatial.html#proper-gaussian-car-model-dcar_proper"><i class="fa fa-check"></i><b>10.2</b> Proper Gaussian CAR model: <em>dcar_proper</em></a>
<ul>
<li class="chapter" data-level="10.2.1" data-path="cha-spatial.html"><a href="cha-spatial.html#specification-and-density-1"><i class="fa fa-check"></i><b>10.2.1</b> Specification and density</a></li>
<li class="chapter" data-level="10.2.2" data-path="cha-spatial.html"><a href="cha-spatial.html#example-1"><i class="fa fa-check"></i><b>10.2.2</b> Example</a></li>
</ul></li>
<li class="chapter" data-level="10.3" data-path="cha-spatial.html"><a href="cha-spatial.html#sec:spatial-mcmc-sampling-car"><i class="fa fa-check"></i><b>10.3</b> MCMC Sampling of CAR models</a>
<ul>
<li class="chapter" data-level="10.3.1" data-path="cha-spatial.html"><a href="cha-spatial.html#initial-values"><i class="fa fa-check"></i><b>10.3.1</b> Initial values</a></li>
<li class="chapter" data-level="10.3.2" data-path="cha-spatial.html"><a href="cha-spatial.html#zero-neighbor-regions"><i class="fa fa-check"></i><b>10.3.2</b> Zero-neighbor regions</a></li>
<li class="chapter" data-level="10.3.3" data-path="cha-spatial.html"><a href="cha-spatial.html#zero-mean-constraint"><i class="fa fa-check"></i><b>10.3.3</b> Zero-mean constraint</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="11" data-path="cha-bnp.html"><a href="cha-bnp.html"><i class="fa fa-check"></i><b>11</b> Bayesian nonparametric models</a>
<ul>
<li class="chapter" data-level="11.1" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:bnpmixtures"><i class="fa fa-check"></i><b>11.1</b> Bayesian nonparametric mixture models</a></li>
<li class="chapter" data-level="11.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:crp"><i class="fa fa-check"></i><b>11.2</b> Chinese Restaurant Process model</a>
<ul>
<li class="chapter" data-level="11.2.1" data-path="cha-bnp.html"><a href="cha-bnp.html#specification-and-density-2"><i class="fa fa-check"></i><b>11.2.1</b> Specification and density</a></li>
<li class="chapter" data-level="11.2.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:excrp"><i class="fa fa-check"></i><b>11.2.2</b> Example</a></li>
<li class="chapter" data-level="11.2.3" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:extensionscrp"><i class="fa fa-check"></i><b>11.2.3</b> Extensions</a></li>
</ul></li>
<li class="chapter" data-level="11.3" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:sb"><i class="fa fa-check"></i><b>11.3</b> Stick-breaking model</a>
<ul>
<li class="chapter" data-level="11.3.1" data-path="cha-bnp.html"><a href="cha-bnp.html#specification-and-function"><i class="fa fa-check"></i><b>11.3.1</b> Specification and function</a></li>
<li class="chapter" data-level="11.3.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:exsb"><i class="fa fa-check"></i><b>11.3.2</b> Example</a></li>
</ul></li>
<li class="chapter" data-level="11.4" data-path="cha-bnp.html"><a href="cha-bnp.html#mcmc-sampling-of-bnp-models"><i class="fa fa-check"></i><b>11.4</b> MCMC sampling of BNP models</a>
<ul>
<li class="chapter" data-level="11.4.1" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:mcmcdcrp"><i class="fa fa-check"></i><b>11.4.1</b> Sampling CRP models</a></li>
<li class="chapter" data-level="11.4.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:mcmcsb"><i class="fa fa-check"></i><b>11.4.2</b> Sampling stick-breaking models</a></li>
</ul></li>
</ul></li>
<li class="part"><span><b>IV Programming with NIMBLE</b></span></li>
<li class="chapter" data-level="" data-path="overview.html"><a href="overview.html"><i class="fa fa-check"></i>Overview</a></li>
<li class="chapter" data-level="12" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html"><i class="fa fa-check"></i><b>12</b> Writing simple nimbleFunctions</a>
<ul>
<li class="chapter" data-level="12.1" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:RC-intro"><i class="fa fa-check"></i><b>12.1</b> Introduction to simple nimbleFunctions</a></li>
<li class="chapter" data-level="12.2" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:r-fiunctions-implemented"><i class="fa fa-check"></i><b>12.2</b> R functions (or variants) implemented in NIMBLE</a>
<ul>
<li class="chapter" data-level="12.2.1" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#finding-help-for-nimbles-versions-of-r-functions"><i class="fa fa-check"></i><b>12.2.1</b> Finding help for NIMBLE’s versions of R functions</a></li>
<li class="chapter" data-level="12.2.2" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#basic-operations"><i class="fa fa-check"></i><b>12.2.2</b> Basic operations</a></li>
<li class="chapter" data-level="12.2.3" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:basic-math-linear"><i class="fa fa-check"></i><b>12.2.3</b> Math and linear algebra</a></li>
<li class="chapter" data-level="12.2.4" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:nimble-dist-funs"><i class="fa fa-check"></i><b>12.2.4</b> Distribution functions</a></li>
<li class="chapter" data-level="12.2.5" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:basic-flow-control"><i class="fa fa-check"></i><b>12.2.5</b> Flow control: <em>if-then-else</em>, <em>for</em>, <em>while</em>, and <em>stop</em></a></li>
<li class="chapter" data-level="12.2.6" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:print"><i class="fa fa-check"></i><b>12.2.6</b> <em>print</em> and <em>cat</em></a></li>
<li class="chapter" data-level="12.2.7" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:check-user-interr"><i class="fa fa-check"></i><b>12.2.7</b> Checking for user interrupts: <em>checkInterrupt</em></a></li>
<li class="chapter" data-level="12.2.8" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#optimization-optim-and-nimoptim"><i class="fa fa-check"></i><b>12.2.8</b> Optimization: <em>optim</em> and <em>nimOptim</em></a></li>
<li class="chapter" data-level="12.2.9" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#integration-integrate-and-nimintegrate"><i class="fa fa-check"></i><b>12.2.9</b> Integration: <em>integrate</em> and <em>nimIntegrate</em></a></li>
<li class="chapter" data-level="12.2.10" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:altern-keyw-some"><i class="fa fa-check"></i><b>12.2.10</b> ‘nim’ synonyms for some functions</a></li>
</ul></li>
<li class="chapter" data-level="12.3" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:how-nimble-handles"><i class="fa fa-check"></i><b>12.3</b> How NIMBLE handles types of variables</a>
<ul>
<li class="chapter" data-level="12.3.1" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:nimbleList-RCFuns"><i class="fa fa-check"></i><b>12.3.1</b> nimbleList data structures</a></li>
<li class="chapter" data-level="12.3.2" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:how-types-work"><i class="fa fa-check"></i><b>12.3.2</b> How numeric types work</a></li>
</ul></li>
<li class="chapter" data-level="12.4" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:decl-argum-return"><i class="fa fa-check"></i><b>12.4</b> Declaring argument and return types</a></li>
<li class="chapter" data-level="12.5" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:comp-nimbl-pass"><i class="fa fa-check"></i><b>12.5</b> Compiled nimbleFunctions pass arguments by reference</a></li>
<li class="chapter" data-level="12.6" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:calling-external-code"><i class="fa fa-check"></i><b>12.6</b> Calling external compiled code</a></li>
<li class="chapter" data-level="12.7" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:calling-R-code"><i class="fa fa-check"></i><b>12.7</b> Calling uncompiled R functions from compiled nimbleFunctions</a></li>
</ul></li>
<li class="chapter" data-level="13" data-path="cha-user-defined.html"><a href="cha-user-defined.html"><i class="fa fa-check"></i><b>13</b> Creating user-defined distributions and functions for models</a>
<ul>
<li class="chapter" data-level="13.1" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:user-functions"><i class="fa fa-check"></i><b>13.1</b> User-defined functions</a></li>
<li class="chapter" data-level="13.2" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:user-distributions"><i class="fa fa-check"></i><b>13.2</b> User-defined distributions</a>
<ul>
<li class="chapter" data-level="13.2.1" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:registerDistributions"><i class="fa fa-check"></i><b>13.2.1</b> Using <em>registerDistributions</em> for alternative parameterizations and providing other information</a></li>
</ul></li>
<li class="chapter" data-level="13.3" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:adv-user-def"><i class="fa fa-check"></i><b>13.3</b> Advanced user-defined functions and distributions</a></li>
<li class="chapter" data-level="13.4" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:user-macros"><i class="fa fa-check"></i><b>13.4</b> User-defined model macros</a></li>
</ul></li>
<li class="chapter" data-level="14" data-path="cha-using-models.html"><a href="cha-using-models.html"><i class="fa fa-check"></i><b>14</b> Working with NIMBLE models</a>
<ul>
<li class="chapter" data-level="14.1" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:accessing-variables"><i class="fa fa-check"></i><b>14.1</b> The variables and nodes in a NIMBLE model</a>
<ul>
<li class="chapter" data-level="14.1.1" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:get-nodes"><i class="fa fa-check"></i><b>14.1.1</b> Determining the nodes in a model</a></li>
<li class="chapter" data-level="14.1.2" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:introduced-nodes"><i class="fa fa-check"></i><b>14.1.2</b> Understanding lifted nodes</a></li>
<li class="chapter" data-level="14.1.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:cdgetdependencies"><i class="fa fa-check"></i><b>14.1.3</b> Determining dependencies in a model</a></li>
</ul></li>
<li class="chapter" data-level="14.2" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:nodeInfo"><i class="fa fa-check"></i><b>14.2</b> Accessing information about nodes and variables</a>
<ul>
<li class="chapter" data-level="14.2.1" data-path="cha-using-models.html"><a href="cha-using-models.html#getting-distributional-information-about-a-node"><i class="fa fa-check"></i><b>14.2.1</b> Getting distributional information about a node</a></li>
<li class="chapter" data-level="14.2.2" data-path="cha-using-models.html"><a href="cha-using-models.html#getting-information-about-a-distribution"><i class="fa fa-check"></i><b>14.2.2</b> Getting information about a distribution</a></li>
<li class="chapter" data-level="14.2.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:getParam"><i class="fa fa-check"></i><b>14.2.3</b> Getting distribution parameter values for a node</a></li>
<li class="chapter" data-level="14.2.4" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:getBound"><i class="fa fa-check"></i><b>14.2.4</b> Getting distribution bounds for a node</a></li>
</ul></li>
<li class="chapter" data-level="14.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:cdcalc-cdsim-cdgetl"><i class="fa fa-check"></i><b>14.3</b> Carrying out model calculations</a>
<ul>
<li class="chapter" data-level="14.3.1" data-path="cha-using-models.html"><a href="cha-using-models.html#core-model-operations-calculation-and-simulation"><i class="fa fa-check"></i><b>14.3.1</b> Core model operations: calculation and simulation</a></li>
<li class="chapter" data-level="14.3.2" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:cdsimn-cdcalcn-cdget"><i class="fa fa-check"></i><b>14.3.2</b> Pre-defined nimbleFunctions for operating on model nodes: <em>simNodes</em>, <em>calcNodes</em>, and <em>getLogProbNodes</em></a></li>
<li class="chapter" data-level="14.3.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:access-log-prob"><i class="fa fa-check"></i><b>14.3.3</b> Accessing log probabilities via <em>logProb</em> variables</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="15" data-path="cha-data-structures.html"><a href="cha-data-structures.html"><i class="fa fa-check"></i><b>15</b> Data structures in NIMBLE</a>
<ul>
<li class="chapter" data-level="15.1" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:modelValues-struct"><i class="fa fa-check"></i><b>15.1</b> The modelValues data structure</a>
<ul>
<li class="chapter" data-level="15.1.1" data-path="cha-data-structures.html"><a href="cha-data-structures.html#creating-modelvalues-objects"><i class="fa fa-check"></i><b>15.1.1</b> Creating modelValues objects</a></li>
<li class="chapter" data-level="15.1.2" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:access-cont-modelv"><i class="fa fa-check"></i><b>15.1.2</b> Accessing contents of modelValues</a></li>
</ul></li>
<li class="chapter" data-level="15.2" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:nimbleLists"><i class="fa fa-check"></i><b>15.2</b> The nimbleList data structure</a>
<ul>
<li class="chapter" data-level="15.2.1" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:predef-nimbleLists"><i class="fa fa-check"></i><b>15.2.1</b> Pre-defined nimbleList types</a></li>
<li class="chapter" data-level="15.2.2" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:eigen-nimFunctions"><i class="fa fa-check"></i><b>15.2.2</b> Using <em>eigen</em> and <em>svd</em> in nimbleFunctions</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="16" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html"><i class="fa fa-check"></i><b>16</b> Writing nimbleFunctions to interact with models</a>
<ul>
<li class="chapter" data-level="16.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:writ-nimble-funct"><i class="fa fa-check"></i><b>16.1</b> Overview</a></li>
<li class="chapter" data-level="16.2" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:using-comp-nimbl"><i class="fa fa-check"></i><b>16.2</b> Using and compiling nimbleFunctions</a></li>
<li class="chapter" data-level="16.3" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#writing-setup-code"><i class="fa fa-check"></i><b>16.3</b> Writing setup code</a>
<ul>
<li class="chapter" data-level="16.3.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#useful-tools-for-setup-functions"><i class="fa fa-check"></i><b>16.3.1</b> Useful tools for setup functions</a></li>
<li class="chapter" data-level="16.3.2" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:access-modify-numer"><i class="fa fa-check"></i><b>16.3.2</b> Accessing and modifying numeric values from setup</a></li>
<li class="chapter" data-level="16.3.3" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#determining-numeric-types-in-nimblefunctions"><i class="fa fa-check"></i><b>16.3.3</b> Determining numeric types in nimbleFunctions</a></li>
<li class="chapter" data-level="16.3.4" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:determ-pers-texttts"><i class="fa fa-check"></i><b>16.3.4</b> Control of setup outputs</a></li>
</ul></li>
<li class="chapter" data-level="16.4" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:nimble-lang-comp"><i class="fa fa-check"></i><b>16.4</b> Writing run code</a>
<ul>
<li class="chapter" data-level="16.4.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:driv-models:-calc"><i class="fa fa-check"></i><b>16.4.1</b> Driving models: <em>calculate</em>, <em>calculateDiff</em>, <em>simulate</em>, <em>getLogProb</em></a></li>
<li class="chapter" data-level="16.4.2" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#getting-and-setting-variable-and-node-values"><i class="fa fa-check"></i><b>16.4.2</b> Getting and setting variable and node values</a></li>
<li class="chapter" data-level="16.4.3" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#getting-parameter-values-and-node-bounds"><i class="fa fa-check"></i><b>16.4.3</b> Getting parameter values and node bounds</a></li>
<li class="chapter" data-level="16.4.4" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:access-model-modelv"><i class="fa fa-check"></i><b>16.4.4</b> Using modelValues objects</a></li>
<li class="chapter" data-level="16.4.5" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:using-model-variable"><i class="fa fa-check"></i><b>16.4.5</b> Using model variables and modelValues in expressions</a></li>
<li class="chapter" data-level="16.4.6" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:incl-other-meth"><i class="fa fa-check"></i><b>16.4.6</b> Including other methods in a nimbleFunction</a></li>
<li class="chapter" data-level="16.4.7" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:using-other-nimbl"><i class="fa fa-check"></i><b>16.4.7</b> Using other nimbleFunctions</a></li>
<li class="chapter" data-level="16.4.8" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:virt-nimbl-nimbl"><i class="fa fa-check"></i><b>16.4.8</b> Virtual nimbleFunctions and nimbleFunctionLists</a></li>
<li class="chapter" data-level="16.4.9" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#character-objects"><i class="fa fa-check"></i><b>16.4.9</b> Character objects</a></li>
<li class="chapter" data-level="16.4.10" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:user-defined-data"><i class="fa fa-check"></i><b>16.4.10</b> User-defined data structures</a></li>
</ul></li>
<li class="chapter" data-level="16.5" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:user-samplers"><i class="fa fa-check"></i><b>16.5</b> Example: writing user-defined samplers to extend NIMBLE’s MCMC engine</a>
<ul>
<li class="chapter" data-level="16.5.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#user-defined-samplers-and-posterior-predictive-nodes"><i class="fa fa-check"></i><b>16.5.1</b> User-defined samplers and posterior predictive nodes</a></li>
</ul></li>
<li class="chapter" data-level="16.6" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#copying-nimblefunctions-and-nimble-models"><i class="fa fa-check"></i><b>16.6</b> Copying nimbleFunctions (and NIMBLE models)</a></li>
<li class="chapter" data-level="16.7" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:debugging"><i class="fa fa-check"></i><b>16.7</b> Debugging nimbleFunctions</a></li>
<li class="chapter" data-level="16.8" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#timing-nimblefunctions-with-run.time"><i class="fa fa-check"></i><b>16.8</b> Timing nimbleFunctions with <em>run.time</em></a></li>
<li class="chapter" data-level="16.9" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#clearing-and-unloading-compiled-objects"><i class="fa fa-check"></i><b>16.9</b> Clearing and unloading compiled objects</a></li>
<li class="chapter" data-level="16.10" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#reducing-memory-usage"><i class="fa fa-check"></i><b>16.10</b> Reducing memory usage</a></li>
</ul></li>
<li class="part"><span><b>V Automatic Derivatives in NIMBLE</b></span></li>
<li class="chapter" data-level="17" data-path="cha-AD.html"><a href="cha-AD.html"><i class="fa fa-check"></i><b>17</b> Automatic Derivatives</a>
<ul>
<li class="chapter" data-level="17.1" data-path="cha-AD.html"><a href="cha-AD.html#sec:use-derivs"><i class="fa fa-check"></i><b>17.1</b> How to turn on derivatives in a model</a></li>
<li class="chapter" data-level="17.2" data-path="cha-AD.html"><a href="cha-AD.html#sec:AD-user-def"><i class="fa fa-check"></i><b>17.2</b> How to support derivatives in user-defined functions and distributions</a></li>
<li class="chapter" data-level="17.3" data-path="cha-AD.html"><a href="cha-AD.html#what-operations-are-and-arent-supported-for-ad"><i class="fa fa-check"></i><b>17.3</b> What operations are and aren’t supported for AD</a></li>
<li class="chapter" data-level="17.4" data-path="cha-AD.html"><a href="cha-AD.html#basics-of-obtaining-derivatives-in-nimblefunctions"><i class="fa fa-check"></i><b>17.4</b> Basics of obtaining derivatives in <code>nimbleFunctions</code></a>
<ul>
<li class="chapter" data-level="17.4.1" data-path="cha-AD.html"><a href="cha-AD.html#checking-derivatives-with-uncompiled-execution"><i class="fa fa-check"></i><b>17.4.1</b> Checking derivatives with uncompiled execution</a></li>
<li class="chapter" data-level="17.4.2" data-path="cha-AD.html"><a href="cha-AD.html#sec:AD-holding-out"><i class="fa fa-check"></i><b>17.4.2</b> Holding some local variables out of derivative tracking</a></li>
<li class="chapter" data-level="17.4.3" data-path="cha-AD.html"><a href="cha-AD.html#sec:AD-multiple-NF"><i class="fa fa-check"></i><b>17.4.3</b> Using AD with multiple nimbleFunctions</a></li>
<li class="chapter" data-level="17.4.4" data-path="cha-AD.html"><a href="cha-AD.html#sec:understanding-more-AD"><i class="fa fa-check"></i><b>17.4.4</b> Understanding more about how AD works: <em>taping</em> of operations</a></li>
<li class="chapter" data-level="17.4.5" data-path="cha-AD.html"><a href="cha-AD.html#resetting-a-nimderivs-call"><i class="fa fa-check"></i><b>17.4.5</b> Resetting a <code>nimDerivs</code> call</a></li>
<li class="chapter" data-level="17.4.6" data-path="cha-AD.html"><a href="cha-AD.html#a-note-on-performance-benchmarking"><i class="fa fa-check"></i><b>17.4.6</b> A note on performance benchmarking</a></li>
</ul></li>
<li class="chapter" data-level="17.5" data-path="cha-AD.html"><a href="cha-AD.html#advanced-uses-double-taping"><i class="fa fa-check"></i><b>17.5</b> Advanced uses: double taping</a></li>
<li class="chapter" data-level="17.6" data-path="cha-AD.html"><a href="cha-AD.html#derivatives-involving-model-calculations"><i class="fa fa-check"></i><b>17.6</b> Derivatives involving model calculations</a>
<ul>
<li class="chapter" data-level="17.6.1" data-path="cha-AD.html"><a href="cha-AD.html#method-1-nimderivs-of-modelcalculate"><i class="fa fa-check"></i><b>17.6.1</b> Method 1: <code>nimDerivs</code> of <code>model$calculate</code></a></li>
<li class="chapter" data-level="17.6.2" data-path="cha-AD.html"><a href="cha-AD.html#method-2-nimderivs-of-a-method-that-calls-modelcalculate"><i class="fa fa-check"></i><b>17.6.2</b> Method 2: <code>nimDerivs</code> of a method that calls <code>model$calculate</code></a></li>
</ul></li>
<li class="chapter" data-level="17.7" data-path="cha-AD.html"><a href="cha-AD.html#sec:parameter-transform"><i class="fa fa-check"></i><b>17.7</b> Parameter transformations</a></li>
</ul></li>
<li class="chapter" data-level="18" data-path="example-maximum-likelihood-estimation-using-optim-with-gradients-from-nimderivs..html"><a href="example-maximum-likelihood-estimation-using-optim-with-gradients-from-nimderivs..html"><i class="fa fa-check"></i><b>18</b> Example: maximum likelihood estimation using <code>optim</code> with gradients from <code>nimDerivs</code>.</a></li>
<li class="chapter" data-level="" data-path="references.html"><a href="references.html"><i class="fa fa-check"></i>References</a></li>
</ul>

      </nav>
    </div>

    <div class="book-body">
      <div class="body-inner">
        <div class="book-header" role="navigation">
          <h1>
            <i class="fa fa-circle-o-notch fa-spin"></i><a href="./"></a>
          </h1>
        </div>

        <div class="page-wrapper" tabindex="-1" role="main">
          <div class="page-inner">

            <section class="normal" id="section-">
<div id="cha-writing-models" class="section level1 hasAnchor" number="5">
<h1><span class="header-section-number">Chapter 5</span> Writing models in NIMBLE’s dialect of BUGS<a href="cha-writing-models.html#cha-writing-models" class="anchor-section" aria-label="Anchor link to header"></a></h1>
<p>Models in NIMBLE are written using a variation on the BUGS language.
From BUGS code, NIMBLE creates a model object. This chapter
describes NIMBLE’s version of BUGS. The next chapter explains how to
build and manipulate model objects.
<!---  With NIMBLE you can also define your own distributions and functions for use in BUGS code; discussion of this functionality is deferred to Chapter \@ref(cha-user-defined) as it requires the use of nimbleFunctions.  --></p>
<div id="sec:supp-feat-bugs" class="section level2 hasAnchor" number="5.1">
<h2><span class="header-section-number">5.1</span> Comparison to BUGS dialects supported by WinBUGS, OpenBUGS and JAGS<a href="cha-writing-models.html#sec:supp-feat-bugs" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<p>Many users will come to NIMBLE with some familiarity with WinBUGS,
OpenBUGS, or JAGS, so we start by summarizing how NIMBLE is similar to
and different from those before documenting NIMBLE’s version of BUGS
more completely. In general, NIMBLE aims to be compatible with the
original BUGS language and also JAGS’ version. However, at this
point, there are some features not supported by NIMBLE, and there are
some extensions that are planned but not implemented.</p>
<div id="supported-features-of-bugs-and-jags" class="section level3 hasAnchor" number="5.1.1">
<h3><span class="header-section-number">5.1.1</span> Supported features of BUGS and JAGS<a href="cha-writing-models.html#supported-features-of-bugs-and-jags" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<ol style="list-style-type: decimal">
<li>Stochastic and deterministic<a href="#fn10" class="footnote-ref" id="fnref10"><sup>10</sup></a> node declarations.</li>
<li>Most univariate and multivariate distributions.</li>
<li>Link functions.</li>
<li>Most mathematical functions.</li>
<li>‘for’ loops for iterative declarations.</li>
<li>Arrays of nodes up to 4 dimensions.</li>
<li>Truncation and censoring as in JAGS using the <code>T()</code>
notation and <code>dinterval</code>.</li>
</ol>
</div>
<div id="sec:extensions-bugs" class="section level3 hasAnchor" number="5.1.2">
<h3><span class="header-section-number">5.1.2</span> NIMBLE’s Extensions to BUGS and JAGS<a href="cha-writing-models.html#sec:extensions-bugs" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>NIMBLE extends the BUGS language in the following ways:</p>
<ol style="list-style-type: decimal">
<li>User-defined functions and distributions – written as nimbleFunctions – can be used in model code. See Chapter <a href="cha-user-defined.html#cha-user-defined">13</a>.</li>
<li>Multiple parameterizations for distributions, similar to those in R, can be used.</li>
<li>Named parameters for distributions and functions, similar to R function calls, can be used.</li>
<li>Linear algebra, including for vectorized
calculations of simple algebra, can be used in deterministic declarations.</li>
<li>Distribution parameters can be expressions, as in JAGS but not in WinBUGS. Caveat: parameters to <em>multivariate</em>
distributions (e.g., <code>dmnorm</code>) cannot be expressions (but an expression can be defined in a separate deterministic expression and the resulting variable then used). <!--- still true. -Perry --></li>
<li>Alternative models can be defined from the same model code by using if-then-else statements that are evaluated when the model is defined.</li>
<li>More flexible indexing of vector nodes within larger variables is allowed. For example one can place a multivariate normal vector arbitrarily within a higher-dimensional object, not just in the last index.</li>
<li>More general constraints can be declared using <code>dconstraint</code>, which extends the concept of JAGS’ <code>dinterval</code>.</li>
<li>Link functions can be used in stochastic, as well as deterministic, declarations.<a href="#fn11" class="footnote-ref" id="fnref11"><sup>11</sup></a></li>
<li>Data values can be reset, and which parts of a model are flagged as data can be changed, allowing one model to be used for different data sets without rebuilding the model each time.</li>
<li>One can use stochastic/dynamic indexes, i.e., indexes can be other nodes or functions of other nodes. For a given dimension of a node being indexed, if the index is not constant, it must be a scalar value. So expressions such as <code>mu[k[i], 3]</code> or <code>mu[k[i], 1:3]</code> or <code>mu[k[i], j[i]]</code> are allowed, but not <code>mu[k[i]:(k[i]+1)]</code><a href="#fn12" class="footnote-ref" id="fnref12"><sup>12</sup></a> Nested dynamic indexes such as <code>mu[k[j[i]]]</code> are also allowed.</li>
</ol>
</div>
<div id="sec:not-yet-supported" class="section level3 hasAnchor" number="5.1.3">
<h3><span class="header-section-number">5.1.3</span> Not-supported features of BUGS and JAGS<a href="cha-writing-models.html#sec:not-yet-supported" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>The following are not supported.</p>
<ol style="list-style-type: decimal">
<li>The appearance of the same node on the left-hand side of both a
<code>&lt;-</code> and a <span class="math inline">\(\sim\)</span> declaration (used in WinBUGS for data
assignment for the value of a stochastic node).</li>
<li>Multivariate nodes must appear with brackets, even if they are
empty. E.g., <code>x</code> cannot be multivariate but <code>x[]</code> or
<code>x[2:5]</code> can be.</li>
<li>NIMBLE generally determines the dimensionality and
sizes of variables from the BUGS code. However, when a variable
appears with blank indices, such as in <code>x.sum &lt;- sum(x[])</code>,
and if the dimensions of the variable are not clearly defined in
other declarations, NIMBLE currently requires that the dimensions of
x be provided when the model object is created (via <code>nimbleModel</code>).</li>
<li>Use of non-sequentially-increasing indexes in the definition of a for loop. JAGS allows syntax such as <code>for(i in expr)</code> when <code>expr</code> evaluates to an integer vector. In NIMBLE one can use the work-around of defining a constant vector, say <code>k</code>, and using <code>k[i]</code> in the body of the for loop.</li>
<li>Use of non-sequentially-increasing indexes to subset variables. JAGS allows syntax such as <code>sum(mu[expr])</code> when <code>expr</code> evaluates to an integer vector. In NIMBLE one can use the work-around of defining a nimbleFunction that takes <code>mu</code> and the index vector as arguments and does the calculation of interest.</li>
</ol>
</div>
</div>
<div id="writing-models" class="section level2 hasAnchor" number="5.2">
<h2><span class="header-section-number">5.2</span> Writing models<a href="cha-writing-models.html#writing-models" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<p>Here we introduce NIMBLE’s version of BUGS. The WinBUGS, OpenBUGS and
JAGS manuals are also useful resources for writing BUGS models,
including many examples.</p>
<div id="declaring-stochastic-and-deterministic-nodes" class="section level3 hasAnchor" number="5.2.1">
<h3><span class="header-section-number">5.2.1</span> Declaring stochastic and deterministic nodes<a href="cha-writing-models.html#declaring-stochastic-and-deterministic-nodes" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>BUGS is a declarative language for graphical (or hierarchical) models.
Most programming languages are imperative, which means a series of
commands will be executed in the order they are written. A
declarative language like BUGS is more like building a machine before
using it. Each line declares that a component should be plugged into
the machine, but it doesn’t matter in what order they are declared as
long as all the right components are plugged in by the end of the code.</p>
<p>The machine in this case is a graphical model<a href="#fn13" class="footnote-ref" id="fnref13"><sup>13</sup></a>. A <em>node</em> (sometimes called
a <em>vertex</em>) holds one value, which may be a scalar or a vector.
<em>Edges</em> define the relationships between nodes. A huge variety
of statistical models can be thought of as graphs.</p>
<p>Here is the code to define and create a simple linear regression model with four observations.</p>
<div class="sourceCode" id="cb77"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb77-1"><a href="cha-writing-models.html#cb77-1" tabindex="-1"></a><span class="fu">library</span>(nimble)</span>
<span id="cb77-2"><a href="cha-writing-models.html#cb77-2" tabindex="-1"></a>mc <span class="ot">&lt;-</span> <span class="fu">nimbleCode</span>({</span>
<span id="cb77-3"><a href="cha-writing-models.html#cb77-3" tabindex="-1"></a>    intercept <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> <span class="dv">1000</span>)</span>
<span id="cb77-4"><a href="cha-writing-models.html#cb77-4" tabindex="-1"></a>    slope <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> <span class="dv">1000</span>)</span>
<span id="cb77-5"><a href="cha-writing-models.html#cb77-5" tabindex="-1"></a>    sigma <span class="sc">~</span> <span class="fu">dunif</span>(<span class="dv">0</span>, <span class="dv">100</span>)</span>
<span id="cb77-6"><a href="cha-writing-models.html#cb77-6" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">4</span>) {</span>
<span id="cb77-7"><a href="cha-writing-models.html#cb77-7" tabindex="-1"></a>        predicted.y[i] <span class="ot">&lt;-</span> intercept <span class="sc">+</span> slope <span class="sc">*</span> x[i]</span>
<span id="cb77-8"><a href="cha-writing-models.html#cb77-8" tabindex="-1"></a>        y[i] <span class="sc">~</span> <span class="fu">dnorm</span>(predicted.y[i], <span class="at">sd =</span> sigma)</span>
<span id="cb77-9"><a href="cha-writing-models.html#cb77-9" tabindex="-1"></a>    }</span>
<span id="cb77-10"><a href="cha-writing-models.html#cb77-10" tabindex="-1"></a>})</span>
<span id="cb77-11"><a href="cha-writing-models.html#cb77-11" tabindex="-1"></a></span>
<span id="cb77-12"><a href="cha-writing-models.html#cb77-12" tabindex="-1"></a>model <span class="ot">&lt;-</span> <span class="fu">nimbleModel</span>(mc, <span class="at">data =</span> <span class="fu">list</span>(<span class="at">y =</span> <span class="fu">rnorm</span>(<span class="dv">4</span>)))</span>
<span id="cb77-13"><a href="cha-writing-models.html#cb77-13" tabindex="-1"></a></span>
<span id="cb77-14"><a href="cha-writing-models.html#cb77-14" tabindex="-1"></a><span class="fu">library</span>(igraph)</span>
<span id="cb77-15"><a href="cha-writing-models.html#cb77-15" tabindex="-1"></a></span>
<span id="cb77-16"><a href="cha-writing-models.html#cb77-16" tabindex="-1"></a>layout <span class="ot">&lt;-</span> <span class="fu">matrix</span>(<span class="at">ncol =</span> <span class="dv">2</span>, <span class="at">byrow =</span> <span class="cn">TRUE</span>,</span>
<span id="cb77-17"><a href="cha-writing-models.html#cb77-17" tabindex="-1"></a>   <span class="co"># These seem to be rescaled to fit in the plot area,</span></span>
<span id="cb77-18"><a href="cha-writing-models.html#cb77-18" tabindex="-1"></a>   <span class="co"># so I&#39;ll just use 0-100 as the scale</span></span>
<span id="cb77-19"><a href="cha-writing-models.html#cb77-19" tabindex="-1"></a>                 <span class="at">data =</span> <span class="fu">c</span>(<span class="dv">33</span>, <span class="dv">100</span>,</span>
<span id="cb77-20"><a href="cha-writing-models.html#cb77-20" tabindex="-1"></a>                          <span class="dv">66</span>, <span class="dv">100</span>,</span>
<span id="cb77-21"><a href="cha-writing-models.html#cb77-21" tabindex="-1"></a>                          <span class="dv">50</span>, <span class="dv">0</span>, <span class="co"># first three are parameters</span></span>
<span id="cb77-22"><a href="cha-writing-models.html#cb77-22" tabindex="-1"></a>                          <span class="dv">15</span>, <span class="dv">50</span>, <span class="dv">35</span>, <span class="dv">50</span>, <span class="dv">55</span>, <span class="dv">50</span>, <span class="dv">75</span>, <span class="dv">50</span>, <span class="co"># x&#39;s</span></span>
<span id="cb77-23"><a href="cha-writing-models.html#cb77-23" tabindex="-1"></a>                          <span class="dv">20</span>, <span class="dv">75</span>, <span class="dv">40</span>, <span class="dv">75</span>, <span class="dv">60</span>, <span class="dv">75</span>, <span class="dv">80</span>, <span class="dv">75</span>, <span class="co"># predicted.y&#39;s</span></span>
<span id="cb77-24"><a href="cha-writing-models.html#cb77-24" tabindex="-1"></a>                          <span class="dv">25</span>, <span class="dv">25</span>, <span class="dv">45</span>, <span class="dv">25</span>, <span class="dv">65</span>, <span class="dv">25</span>, <span class="dv">85</span>, <span class="dv">25</span>) <span class="co"># y&#39;s</span></span>
<span id="cb77-25"><a href="cha-writing-models.html#cb77-25" tabindex="-1"></a>                 )</span>
<span id="cb77-26"><a href="cha-writing-models.html#cb77-26" tabindex="-1"></a></span>
<span id="cb77-27"><a href="cha-writing-models.html#cb77-27" tabindex="-1"></a>sizes <span class="ot">&lt;-</span> <span class="fu">c</span>(<span class="dv">45</span>, <span class="dv">30</span>, <span class="dv">30</span>,</span>
<span id="cb77-28"><a href="cha-writing-models.html#cb77-28" tabindex="-1"></a>           <span class="fu">rep</span>(<span class="dv">20</span>, <span class="dv">4</span>),</span>
<span id="cb77-29"><a href="cha-writing-models.html#cb77-29" tabindex="-1"></a>           <span class="fu">rep</span>(<span class="dv">50</span>, <span class="dv">4</span>),</span>
<span id="cb77-30"><a href="cha-writing-models.html#cb77-30" tabindex="-1"></a>           <span class="fu">rep</span>(<span class="dv">20</span>, <span class="dv">4</span>))</span>
<span id="cb77-31"><a href="cha-writing-models.html#cb77-31" tabindex="-1"></a></span>
<span id="cb77-32"><a href="cha-writing-models.html#cb77-32" tabindex="-1"></a>edge.color <span class="ot">&lt;-</span> <span class="st">&quot;black&quot;</span></span>
<span id="cb77-33"><a href="cha-writing-models.html#cb77-33" tabindex="-1"></a>    <span class="co"># c(</span></span>
<span id="cb77-34"><a href="cha-writing-models.html#cb77-34" tabindex="-1"></a>    <span class="co"># rep(&quot;green&quot;, 8),</span></span>
<span id="cb77-35"><a href="cha-writing-models.html#cb77-35" tabindex="-1"></a>    <span class="co"># rep(&quot;red&quot;, 4),</span></span>
<span id="cb77-36"><a href="cha-writing-models.html#cb77-36" tabindex="-1"></a>    <span class="co"># rep(&quot;blue&quot;, 4),</span></span>
<span id="cb77-37"><a href="cha-writing-models.html#cb77-37" tabindex="-1"></a>    <span class="co"># rep(&quot;purple&quot;, 4))</span></span>
<span id="cb77-38"><a href="cha-writing-models.html#cb77-38" tabindex="-1"></a>stoch.color <span class="ot">&lt;-</span> <span class="st">&quot;deepskyblue2&quot;</span></span>
<span id="cb77-39"><a href="cha-writing-models.html#cb77-39" tabindex="-1"></a>det.color <span class="ot">&lt;-</span> <span class="st">&quot;orchid3&quot;</span></span>
<span id="cb77-40"><a href="cha-writing-models.html#cb77-40" tabindex="-1"></a>rhs.color <span class="ot">&lt;-</span> <span class="st">&quot;gray73&quot;</span></span>
<span id="cb77-41"><a href="cha-writing-models.html#cb77-41" tabindex="-1"></a>fill.color <span class="ot">&lt;-</span> <span class="fu">c</span>(</span>
<span id="cb77-42"><a href="cha-writing-models.html#cb77-42" tabindex="-1"></a>    <span class="fu">rep</span>(stoch.color, <span class="dv">3</span>),</span>
<span id="cb77-43"><a href="cha-writing-models.html#cb77-43" tabindex="-1"></a>    <span class="fu">rep</span>(rhs.color, <span class="dv">4</span>),</span>
<span id="cb77-44"><a href="cha-writing-models.html#cb77-44" tabindex="-1"></a>    <span class="fu">rep</span>(det.color, <span class="dv">4</span>),</span>
<span id="cb77-45"><a href="cha-writing-models.html#cb77-45" tabindex="-1"></a>    <span class="fu">rep</span>(stoch.color, <span class="dv">4</span>)</span>
<span id="cb77-46"><a href="cha-writing-models.html#cb77-46" tabindex="-1"></a>)</span>
<span id="cb77-47"><a href="cha-writing-models.html#cb77-47" tabindex="-1"></a></span>
<span id="cb77-48"><a href="cha-writing-models.html#cb77-48" tabindex="-1"></a></span>
<span id="cb77-49"><a href="cha-writing-models.html#cb77-49" tabindex="-1"></a><span class="fu">plot</span>(model<span class="sc">$</span>graph, <span class="at">vertex.shape =</span> <span class="st">&quot;crectangle&quot;</span>,</span>
<span id="cb77-50"><a href="cha-writing-models.html#cb77-50" tabindex="-1"></a>     <span class="at">vertex.size =</span> sizes,</span>
<span id="cb77-51"><a href="cha-writing-models.html#cb77-51" tabindex="-1"></a>     <span class="at">vertex.size2 =</span> <span class="dv">20</span>,</span>
<span id="cb77-52"><a href="cha-writing-models.html#cb77-52" tabindex="-1"></a>     <span class="at">layout =</span> layout,</span>
<span id="cb77-53"><a href="cha-writing-models.html#cb77-53" tabindex="-1"></a>     <span class="at">vertex.label.cex =</span> <span class="fl">3.0</span>,</span>
<span id="cb77-54"><a href="cha-writing-models.html#cb77-54" tabindex="-1"></a>     <span class="at">vertex.color =</span> fill.color,</span>
<span id="cb77-55"><a href="cha-writing-models.html#cb77-55" tabindex="-1"></a>     <span class="at">edge.width =</span> <span class="dv">3</span>,</span>
<span id="cb77-56"><a href="cha-writing-models.html#cb77-56" tabindex="-1"></a>     <span class="at">asp =</span> <span class="fl">0.5</span>,</span>
<span id="cb77-57"><a href="cha-writing-models.html#cb77-57" tabindex="-1"></a>     <span class="at">edge.color =</span> edge.color)</span></code></pre></div>
<div class="figure"><span style="display:block;" id="fig:linearRegressionGraph"></span>
<img src="NimbleUserManual_files/figure-html/linearRegressionGraph-1.png" alt="Graph of a linear regression model" width="1920" />
<p class="caption">
Figure 5.1: Graph of a linear regression model
</p>
</div>
<p>The graph representing the model is
shown in Figure <a href="cha-writing-models.html#fig:linearRegressionGraph">5.1</a>. Each observation,
<code>y[i]</code>, is a node whose edges say that it follows a normal
distribution depending on a predicted value, <code>predicted.y[i]</code>, and
standard deviation, <code>sigma</code>, which are each nodes. Each predicted
value is a node whose edges say how it is calculated from <code>slope</code>,
<code>intercept</code>, and one value of an explanatory variable, <code>x[i]</code>,
which are each nodes.</p>
<p>This graph is created from the following BUGS code:</p>
<div class="sourceCode" id="cb78"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb78-1"><a href="cha-writing-models.html#cb78-1" tabindex="-1"></a>{</span>
<span id="cb78-2"><a href="cha-writing-models.html#cb78-2" tabindex="-1"></a>    intercept <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> <span class="dv">1000</span>)</span>
<span id="cb78-3"><a href="cha-writing-models.html#cb78-3" tabindex="-1"></a>    slope <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> <span class="dv">1000</span>)</span>
<span id="cb78-4"><a href="cha-writing-models.html#cb78-4" tabindex="-1"></a>    sigma <span class="sc">~</span> <span class="fu">dunif</span>(<span class="dv">0</span>, <span class="dv">100</span>)</span>
<span id="cb78-5"><a href="cha-writing-models.html#cb78-5" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">4</span>) {</span>
<span id="cb78-6"><a href="cha-writing-models.html#cb78-6" tabindex="-1"></a>        predicted.y[i] <span class="ot">&lt;-</span> intercept <span class="sc">+</span> slope <span class="sc">*</span> x[i]</span>
<span id="cb78-7"><a href="cha-writing-models.html#cb78-7" tabindex="-1"></a>        y[i] <span class="sc">~</span> <span class="fu">dnorm</span>(predicted.y[i], <span class="at">sd =</span> sigma)</span>
<span id="cb78-8"><a href="cha-writing-models.html#cb78-8" tabindex="-1"></a>    }</span>
<span id="cb78-9"><a href="cha-writing-models.html#cb78-9" tabindex="-1"></a>}</span></code></pre></div>
<p>In this code, stochastic relationships are declared with ‘<span class="math inline">\(\sim\)</span>’
and deterministic relationships are declared with ‘<code>&lt;-</code>’. For
example, each <code>y[i]</code> follows a normal distribution with mean
<code>predicted.y[i]</code> and standard deviation
<code>sigma</code>. Each
<code>predicted.y[i]</code> is the result of <code>intercept + slope * x[i]</code>.
The for-loop yields the equivalent of writing four lines of code, each
with a different value of <code>i</code>. It does not matter in what order
the nodes are declared. Imagine that each line of code draws part of
Figure <a href="cha-writing-models.html#fig:linearRegressionGraph">5.1</a>, and all that matters is that
the everything gets drawn in the end. Available distributions, default and alternative
parameterizations, and functions are listed in Section <a href="cha-writing-models.html#subsec:dists-and-functions">5.2.4</a>.</p>
<p>An equivalent graph can be created by this BUGS code:</p>
<div class="sourceCode" id="cb79"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb79-1"><a href="cha-writing-models.html#cb79-1" tabindex="-1"></a>{</span>
<span id="cb79-2"><a href="cha-writing-models.html#cb79-2" tabindex="-1"></a>    intercept <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> <span class="dv">1000</span>)</span>
<span id="cb79-3"><a href="cha-writing-models.html#cb79-3" tabindex="-1"></a>    slope <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> <span class="dv">1000</span>)</span>
<span id="cb79-4"><a href="cha-writing-models.html#cb79-4" tabindex="-1"></a>    sigma <span class="sc">~</span> <span class="fu">dunif</span>(<span class="dv">0</span>, <span class="dv">100</span>)</span>
<span id="cb79-5"><a href="cha-writing-models.html#cb79-5" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">4</span>) {</span>
<span id="cb79-6"><a href="cha-writing-models.html#cb79-6" tabindex="-1"></a>        y[i] <span class="sc">~</span> <span class="fu">dnorm</span>(intercept <span class="sc">+</span> slope <span class="sc">*</span> x[i], <span class="at">sd =</span> sigma)</span>
<span id="cb79-7"><a href="cha-writing-models.html#cb79-7" tabindex="-1"></a>    }</span>
<span id="cb79-8"><a href="cha-writing-models.html#cb79-8" tabindex="-1"></a>}</span></code></pre></div>
<p>In this case, the <code>predicted.y[i]</code> nodes in Figure
<a href="cha-writing-models.html#fig:linearRegressionGraph">5.1</a> will be created automatically by
NIMBLE and will have a different name, generated by NIMBLE.</p>
</div>
<div id="sec:more-kinds-bugs" class="section level3 hasAnchor" number="5.2.2">
<h3><span class="header-section-number">5.2.2</span> More kinds of BUGS declarations<a href="cha-writing-models.html#sec:more-kinds-bugs" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>Here are some examples of valid lines of BUGS code. This code does
not describe a sensible or complete model, and it includes some
arbitrary indices (e.g. <code>mvx[8:10, i]</code>) to illustrate flexibility.
Instead the purpose of each line is to illustrate a feature of
NIMBLE’s version of BUGS.</p>
<div class="sourceCode" id="cb80"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb80-1"><a href="cha-writing-models.html#cb80-1" tabindex="-1"></a>{</span>
<span id="cb80-2"><a href="cha-writing-models.html#cb80-2" tabindex="-1"></a>    <span class="co"># 1. normal distribution with BUGS parameter order</span></span>
<span id="cb80-3"><a href="cha-writing-models.html#cb80-3" tabindex="-1"></a>    x <span class="sc">~</span> <span class="fu">dnorm</span>(a <span class="sc">+</span> b <span class="sc">*</span> c, tau) </span>
<span id="cb80-4"><a href="cha-writing-models.html#cb80-4" tabindex="-1"></a>    <span class="co"># 2. normal distribution with a named parameter</span></span>
<span id="cb80-5"><a href="cha-writing-models.html#cb80-5" tabindex="-1"></a>    y <span class="sc">~</span> <span class="fu">dnorm</span>(a <span class="sc">+</span> b <span class="sc">*</span> c, <span class="at">sd =</span> sigma) </span>
<span id="cb80-6"><a href="cha-writing-models.html#cb80-6" tabindex="-1"></a>    <span class="co"># 3. For-loop and nested indexing</span></span>
<span id="cb80-7"><a href="cha-writing-models.html#cb80-7" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span>N) {</span>
<span id="cb80-8"><a href="cha-writing-models.html#cb80-8" tabindex="-1"></a>        <span class="cf">for</span>(j <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span>M[i]) {</span>
<span id="cb80-9"><a href="cha-writing-models.html#cb80-9" tabindex="-1"></a>            z[i,j] <span class="sc">~</span> <span class="fu">dexp</span>(r[ blockID[i] ]) </span>
<span id="cb80-10"><a href="cha-writing-models.html#cb80-10" tabindex="-1"></a>        }</span>
<span id="cb80-11"><a href="cha-writing-models.html#cb80-11" tabindex="-1"></a>    }</span>
<span id="cb80-12"><a href="cha-writing-models.html#cb80-12" tabindex="-1"></a>    <span class="co"># 4. multivariate distribution with arbitrary indexing</span></span>
<span id="cb80-13"><a href="cha-writing-models.html#cb80-13" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>) </span>
<span id="cb80-14"><a href="cha-writing-models.html#cb80-14" tabindex="-1"></a>        mvx[<span class="dv">8</span><span class="sc">:</span><span class="dv">10</span>, i] <span class="sc">~</span> <span class="fu">dmnorm</span>(mvMean[<span class="dv">3</span><span class="sc">:</span><span class="dv">5</span>], <span class="at">cov =</span> mvCov[<span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>, <span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>, i])</span>
<span id="cb80-15"><a href="cha-writing-models.html#cb80-15" tabindex="-1"></a>    <span class="co"># 5. User-provided distribution</span></span>
<span id="cb80-16"><a href="cha-writing-models.html#cb80-16" tabindex="-1"></a>    w <span class="sc">~</span> <span class="fu">dMyDistribution</span>(<span class="at">hello =</span> x, <span class="at">world =</span> y) </span>
<span id="cb80-17"><a href="cha-writing-models.html#cb80-17" tabindex="-1"></a>    <span class="co"># 6. Simple deterministic node</span></span>
<span id="cb80-18"><a href="cha-writing-models.html#cb80-18" tabindex="-1"></a>    d1 <span class="ot">&lt;-</span> a <span class="sc">+</span> b</span>
<span id="cb80-19"><a href="cha-writing-models.html#cb80-19" tabindex="-1"></a>    <span class="co"># 7. Vector deterministic node with matrix multiplication</span></span>
<span id="cb80-20"><a href="cha-writing-models.html#cb80-20" tabindex="-1"></a>    d2[] <span class="ot">&lt;-</span> A[ , ] <span class="sc">%*%</span> mvMean[<span class="dv">1</span><span class="sc">:</span><span class="dv">5</span>] </span>
<span id="cb80-21"><a href="cha-writing-models.html#cb80-21" tabindex="-1"></a>    <span class="co"># 8. Deterministic node with user-provided function</span></span>
<span id="cb80-22"><a href="cha-writing-models.html#cb80-22" tabindex="-1"></a>    d3 <span class="ot">&lt;-</span> <span class="fu">foo</span>(x, <span class="at">hooray =</span> y) </span>
<span id="cb80-23"><a href="cha-writing-models.html#cb80-23" tabindex="-1"></a>}</span></code></pre></div>
<p>When a variable appears only on the right-hand side, it can be provided via <code>constants</code> (in which case it can never be changed) or via <code>data</code> or <code>inits</code>, as discussed in Chapter <a href="cha-building-models.html#cha-building-models">6</a>.</p>
<p>Notes on the comment-numbered lines are:</p>
<ol style="list-style-type: decimal">
<li><code>x</code> follows a normal distribution with mean <code>a + b*c</code> and precision <code>tau</code> (default BUGS second parameter for <code>dnorm</code>).</li>
<li><code>y</code> follows a normal distribution with the same mean as <code>x</code> but a named standard deviation parameter instead of a precision parameter (sd = 1/sqrt(precision)).</li>
<li><code>z[i, j]</code> follows an exponential distribution with parameter
<code>r[ blockID[i] ]</code>. This shows how for-loops can be used for indexing of variables containing
multiple nodes. Variables that define for-loop indices (<code>N</code> and <code>M</code>) must also be provided as constants.<br />
</li>
<li>The arbitrary block <code>mvx[8:10, i]</code> follows a multivariate
normal distribution, with a named covariance matrix instead of BUGS’
default of a precision matrix. As in R, curly braces for for-loop
contents are only needed if there is more than one line.</li>
<li><code>w</code> follows a user-defined distribution. See Chapter <a href="cha-user-defined.html#cha-user-defined">13</a>.</li>
<li><code>d1</code> is a scalar deterministic node that, when calculated, will be
set to <code>a + b</code>.</li>
<li><code>d2</code> is a vector deterministic node using matrix
multiplication in R’s syntax.</li>
<li><code>d3</code> is a deterministic node using a user-provided
function. See Chapter <a href="cha-user-defined.html#cha-user-defined">13</a>.</li>
</ol>
<div id="sec:indexing" class="section level4 hasAnchor" number="5.2.2.1">
<h4><span class="header-section-number">5.2.2.1</span> More about indexing<a href="cha-writing-models.html#sec:indexing" class="anchor-section" aria-label="Anchor link to header"></a></h4>
<p>Examples of allowed indexing include:</p>
<ul>
<li><code>x[i]</code> # a single index</li>
<li><code>x[i:j]</code> # a range of indices</li>
<li><code>x[i:j,k:l]</code> # multiple single indices or ranges for higher-dimensional arrays</li>
<li><code>x[i:j, ]</code> # blank indices indicating the full range</li>
<li><code>x[3*i+7]</code> # computed indices</li>
<li><code>x[(3*i):(5*i+1)]</code> # computed lower and upper ends of an index range</li>
<li><code>x[k[i]+1]</code> # a dynamic (and computed) index</li>
<li><code>x[k[j[i]]]</code> # nested dynamic indexes</li>
<li><code>x[k[i], 1:3]</code> # nested indexing of rows or columns</li>
</ul>
<p>NIMBLE does not allow multivariate nodes to be used without
square brackets, which is an incompatibility with JAGS. Therefore a statement like <code>xbar &lt;- mean(x)</code> in JAGS must be converted to
<code>xbar &lt;- mean(x[])</code> (if <code>x</code> is a vector) or <code>xbar &lt;-  mean(x[,])</code> (if <code>x</code> is a matrix) for NIMBLE<a href="#fn14" class="footnote-ref" id="fnref14"><sup>14</sup></a>. Section <a href="cha-building-models.html#sec:provide-dimensions">6.1.1.5</a> discusses how to provide NIMBLE with dimensions of <code>x</code> when needed.</p>
<p>One cannot provide a vector of indices that are not constant. For example, <code>x[start[i]:end[i]]</code> is allowed only if <code>start</code> and <code>end</code> are provided as constants. Also, one cannot provide a vector, or expression evaluating to a vector (apart from use of <code>:</code>) as an index. For example, <code>x[inds]</code> is not allowed. Often one can write a nimbleFunction to achieve the desired result, such as definining a nimbleFunction that takes <code>x</code>, <code>start[i]</code>, and <code>end[i]</code> or takes <code>x</code> and <code>inds</code> as arguments and does the subsetting (possibly in combination with some other calculation). Note that <code>x[L:U]</code> and <code>x[inds]</code> are allowed in nimbleFunction code.</p>
<p>Generally NIMBLE supports R-like linear algebra expressions and attempts to follow the same rules as R about
dimensions (although in some cases this is not possible). For
example, <code>x[1:3] %*% y[1:3]</code> converts <code>x[1:3]</code> into a row
vector and thus computes the inner product, which is returned as a <span class="math inline">\(1 \times 1\)</span>
matrix (use <code>inprod</code> to get it as a scalar, which it typically easier). Like in R,
a scalar index will result in dropping a dimension unless the argument
<code>drop=FALSE</code> is provided. For example, <code>mymatrix[i, 1:3]</code> will
be a vector of length 3, but <code>mymatrix[i, 1:3, drop=FALSE]</code> will be
a <span class="math inline">\(1 \times 3\)</span> matrix. More about indexing and dimensions is
discussed in Section <a href="cha-RCfunctions.html#sec:manag-dimens-sizes">12.3.2.7</a>.</p>
</div>
</div>
<div id="subsec:vectorized-versus-scalar-declarations" class="section level3 hasAnchor" number="5.2.3">
<h3><span class="header-section-number">5.2.3</span> Vectorized versus scalar declarations<a href="cha-writing-models.html#subsec:vectorized-versus-scalar-declarations" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>Suppose you need nodes <code>logY[i]</code> that should be the log of the
corresponding <code>Y[i]</code>, say for <code>i</code> from 1 to 10. Conventionally
this would be created with a for loop:</p>
<div class="sourceCode" id="cb81"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb81-1"><a href="cha-writing-models.html#cb81-1" tabindex="-1"></a>{</span>
<span id="cb81-2"><a href="cha-writing-models.html#cb81-2" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">10</span>) {</span>
<span id="cb81-3"><a href="cha-writing-models.html#cb81-3" tabindex="-1"></a>        logY[i] <span class="ot">&lt;-</span> <span class="fu">log</span>(Y[i])</span>
<span id="cb81-4"><a href="cha-writing-models.html#cb81-4" tabindex="-1"></a>    }</span>
<span id="cb81-5"><a href="cha-writing-models.html#cb81-5" tabindex="-1"></a>}</span></code></pre></div>
<p>Since NIMBLE supports R-like algebraic expressions, an alternative in
NIMBLE’s dialect of BUGS is to use a vectorized declaration like this:</p>
<div class="sourceCode" id="cb82"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb82-1"><a href="cha-writing-models.html#cb82-1" tabindex="-1"></a>{</span>
<span id="cb82-2"><a href="cha-writing-models.html#cb82-2" tabindex="-1"></a>    logY[<span class="dv">1</span><span class="sc">:</span><span class="dv">10</span>] <span class="ot">&lt;-</span> <span class="fu">log</span>(Y[<span class="dv">1</span><span class="sc">:</span><span class="dv">10</span>])</span>
<span id="cb82-3"><a href="cha-writing-models.html#cb82-3" tabindex="-1"></a>}</span></code></pre></div>
<p>There is an important difference between the models that are created by the
above two methods. The first creates 10 scalar nodes, <code>logY[1]</code>
<span class="math inline">\(,\ldots,\)</span> <code>logY[10]</code>. The second creates one vector node,
<code>logY[1:10]</code>. If each <code>logY[i]</code> is used separately by an algorithm, it may be more efficient computationally if they are declared as scalars. If they are all used together,
it will often make sense to declare them as a vector.</p>
</div>
<div id="subsec:dists-and-functions" class="section level3 hasAnchor" number="5.2.4">
<h3><span class="header-section-number">5.2.4</span> Available distributions<a href="cha-writing-models.html#subsec:dists-and-functions" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<div id="subsec:distributions" class="section level4 hasAnchor" number="5.2.4.1">
<h4><span class="header-section-number">5.2.4.1</span> Distributions<a href="cha-writing-models.html#subsec:distributions" class="anchor-section" aria-label="Anchor link to header"></a></h4>
<p>NIMBLE supports most of the distributions allowed in BUGS and
JAGS. Table <a href="cha-writing-models.html#tab:distributions">5.1</a> lists the distributions that are
currently supported, with their default parameterizations, which match
those of BUGS<a href="#fn15" class="footnote-ref" id="fnref15"><sup>15</sup></a>. NIMBLE also allows one to use alternative
parameterizations for a variety of distributions as described next.
See Section <a href="cha-user-defined.html#sec:user-distributions">13.2</a> to learn how to write new distributions using nimbleFunctions.</p>
<table>
<caption><span id="tab:distributions">Table 5.1: </span> Distributions with their default order of parameters. The value of the random variable
is given by <span class="math inline">\(x\)</span>.</caption>
<thead>
<tr class="header">
<th align="left">Name</th>
<th align="left">Usage</th>
<th align="left">Density</th>
<th>Lower</th>
<th>Upper</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td align="left">Bernoulli</td>
<td align="left"><code>dbern(prob = p)</code></td>
<td align="left"><span class="math inline">\(p^x (1 - p)^{1 -x}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td><span class="math inline">\(1\)</span></td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left"><span class="math inline">\(0 &lt; p &lt; 1\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">Beta</td>
<td align="left"><code>dbeta(shape1 = a,</code></td>
<td align="left"><span class="math inline">\(\frac{x^{a-1}(1-x)^{b-1}}{\beta(a,b)}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td><span class="math inline">\(1\)</span></td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left"><code>shape2 = b)</code>, <span class="math inline">\(a &gt; 0\)</span>, <span class="math inline">\(b &gt; 0\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">Binomial</td>
<td align="left"><code>dbin(prob = p, size = n)</code></td>
<td align="left"><span class="math inline">\({n \choose x} p^x (1-p)^{n-x}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td><span class="math inline">\(n\)</span></td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left"><span class="math inline">\(0 &lt; p &lt; 1\)</span>, <span class="math inline">\(n \in \mathbb{N}^*\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">CAR</td>
<td align="left"><code>dcar_normal(adj, weights,</code></td>
<td align="left">see chapter <a href="cha-spatial.html#cha-spatial">10</a> for details</td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">(intrinsic)</td>
<td align="left"><code>num, tau, c, zero_mean)</code></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">CAR</td>
<td align="left"><code>dcar_proper(mu, C, adj,</code></td>
<td align="left">see chapter <a href="cha-spatial.html#cha-spatial">10</a> for details</td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">(proper)</td>
<td align="left"><code>num, M, tau, gamma)</code></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">Categorical</td>
<td align="left"><code>dcat(prob = p)</code></td>
<td align="left"><span class="math inline">\(\frac{p_x}{\sum_i p_i}\)</span></td>
<td><span class="math inline">\(1\)</span></td>
<td><span class="math inline">\(N\)</span></td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left"><span class="math inline">\(p \in (\mathbb{R}^+)^N\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">Chi-square</td>
<td align="left"><code>dchisq(df = k)</code>, <span class="math inline">\(k &gt; 0\)</span></td>
<td align="left"><span class="math inline">\(\frac{x^{\frac{k}{2}-1} \exp(-x/2)}{2^{\frac{k}{2}} \Gamma(\frac{k}{2})}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td></td>
</tr>
<tr class="even">
<td align="left">Double</td>
<td align="left"><code>ddexp(location =</code> <span class="math inline">\(\mu\)</span>,</td>
<td align="left"><span class="math inline">\(\frac{\tau}{2} \exp(-\tau |x - \mu|)\)</span></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">exponential</td>
<td align="left"><code>rate =</code> <span class="math inline">\(\tau\)</span><code>),</code><span class="math inline">\(\tau &gt; 0\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">Dirichlet</td>
<td align="left"><code>ddirch(alpha =</code> <span class="math inline">\(\alpha\)</span><code>)</code></td>
<td align="left"><span class="math inline">\(\Gamma(\sum_j \alpha_j) \prod_j \frac{ x_j^{\alpha_j - 1}}{ \Gamma(\alpha_j)}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td></td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left"><span class="math inline">\(\alpha_j \geq 0\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">Exponential</td>
<td align="left"><code>dexp(rate =</code> <span class="math inline">\(\lambda\)</span><code>)</code>, <span class="math inline">\(\lambda &gt; 0\)</span></td>
<td align="left"><span class="math inline">\(\lambda \exp(-\lambda x)\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">Flat</td>
<td align="left"><code>dflat()</code></td>
<td align="left"><span class="math inline">\(\propto 1\)</span> (improper)</td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">Gamma</td>
<td align="left"><code>dgamma(shape = r, rate =</code> <span class="math inline">\(\lambda\)</span><code>)</code></td>
<td align="left"><span class="math inline">\(\frac{ \lambda^r x^{r - 1} \exp(-\lambda x)}{ \Gamma(r)}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td></td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left"><span class="math inline">\(\lambda &gt; 0\)</span>, <span class="math inline">\(r &gt; 0\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">Half flat</td>
<td align="left"><code>dhalfflat()</code></td>
<td align="left"><span class="math inline">\(\propto 1\)</span> (improper)</td>
<td><span class="math inline">\(0\)</span></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">Inverse</td>
<td align="left"><code>dinvgamma(shape = r, scale =</code> <span class="math inline">\(\lambda\)</span><code>)</code></td>
<td align="left"><span class="math inline">\(\frac{ \lambda^r x^{-(r + 1)} \exp(-\lambda / x)}{ \Gamma(r)}\)</span></td>
<td><span class="math inline">\(0\)</span></td>
<td></td>
</tr>
<tr class="even">
<td align="left">gamma</td>
<td align="left"><span class="math inline">\(\lambda &gt; 0\)</span>, <span class="math inline">\(r &gt; 0\)</span></td>
<td align="left"></td>
<td></td>
<td></td>
</tr>
<tr class="odd">
<td align="left">LKJ Correl’n</td>
<td align="left"><code>dlkj_corr_cholesky(eta =</code> <span class="math inline">\(\eta\)</span><code>,</code></td>
<td align="left"><span class="math inline">\(\prod_{i=2}^{p} x_{ii}^{p - i + 2\eta -2}\)</span></td>
<td></td>
<td></td>
</tr>
<tr class="even">
<td align="left">Cholesky</td>
<td align="left"><code>size = p),</code> <span class="math inline">\(\eta &gt; 0\)</span></td>
<td align="left"></td>
<td></td>