-
Notifications
You must be signed in to change notification settings - Fork 16
Expand file tree
/
Copy pathAlgorithmic-Fairness-in-R.html
More file actions
1678 lines (1471 loc) · 65.6 KB
/
Algorithmic-Fairness-in-R.html
File metadata and controls
1678 lines (1471 loc) · 65.6 KB
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="en">
<head>
<title>Algorithmic Fairness in R: fairml & aif360 for Bias Auditing</title>
<meta charset="utf-8">
<meta name="Description" content="Audit machine learning models for fairness using R. Learn disparate impact, calibration, and practical bias auditing with fairml and aif360.">
<meta name="Keywords" content="algorithmic fairness R, fairml, aif360 R, bias auditing, disparate impact, model fairness, fairness metrics R">
<meta name="Distribution" content="Global">
<meta name="Author" content="Selva Prabhakaran">
<meta name="Robots" content="index, follow">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="shortcut icon" href="/screenshots/iconb-64.png" type="image/x-icon" />
<link rel="canonical" href="https://r-statistics.co/Algorithmic-Fairness-in-R.html">
<link rel="alternate" type="application/atom+xml" title="r-statistics.co" href="https://r-statistics.co/feed.xml">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link rel="preconnect" href="https://cdnjs.cloudflare.com">
<!-- Critical CSS inlined for fast first paint -->
<style>
*,*::before,*::after{box-sizing:border-box}
body{margin:0;font-family:'IBM Plex Sans',-apple-system,BlinkMacSystemFont,sans-serif;font-size:18px;line-height:1.7;color:#1a1a2e;background:#fdfdfd}
.container{max-width:1170px;margin:0 auto;padding:0 15px}
.row{margin-left:-15px;margin-right:-15px}.row::after{content:"";display:table;clear:both}
.col-xs-12,.col-sm-2,.col-sm-3,.col-sm-7{position:relative;min-height:1px;padding-left:15px;padding-right:15px;float:left}
.col-xs-12{width:100%}
@media(min-width:768px){.col-sm-2{width:16.667%}.col-sm-3{width:25%}.col-sm-7{width:58.333%}.hidden-xs{display:block!important}}
.hidden-xs{display:none}
.table{width:100%;border-collapse:collapse}.table>thead>tr>th,.table>tbody>tr>td{padding:8px;border-top:1px solid #ddd}
.table-striped>tbody>tr:nth-of-type(odd){background:#f9f9f9}
.btn{display:inline-block;padding:6px 12px;font-size:14px;border-radius:4px;cursor:pointer;border:1px solid transparent}
.btn-primary{color:#fff;background:#3f73d8;border-color:#3f73d8}.btn-primary:hover{background:#2c5fbe}
.btn-default{color:#333;background:#fff;border-color:#ccc}
.btn-sm{padding:3px 10px;font-size:12px}
.form-control{display:block;padding:6px 12px;font-size:14px;border:1px solid #ccc;border-radius:4px}
.list-unstyled{list-style:none;padding-left:0}
.img-responsive{max-width:100%;height:auto}
.pull-right{float:right}
a{color:#3F73D8;text-decoration:none}a:hover{text-decoration:underline}
html,body{overflow-x:hidden;max-width:100vw}
</style>
<!-- Full Bootstrap deferred (non-render-blocking) -->
<link href="www/bootstrap.min.css" rel="stylesheet" media="print" onload="this.media='all'">
<noscript><link href="www/bootstrap.min.css" rel="stylesheet"></noscript>
<link href="www/highlight.css" rel="stylesheet" media="print" onload="this.media='all'">
<link href="css/main.css?v=18" rel="stylesheet">
<link href="https://fonts.googleapis.com/css2?family=IBM+Plex+Sans:wght@400;500;600;700&family=IBM+Plex+Serif:wght@600;700&family=IBM+Plex+Mono:wght@400;500&display=swap"
rel="stylesheet" media="print" onload="this.media='all'">
<noscript><link href="https://fonts.googleapis.com/css2?family=IBM+Plex+Sans:wght@400;500;600;700&family=IBM+Plex+Serif:wght@600;700&family=IBM+Plex+Mono:wght@400;500&display=swap" rel="stylesheet"></noscript>
<!-- Color Script -->
<style type="text/css">
a {
color: #3F73D8;
}
li {
line-height: 1.65;
}
/* reduce spacing around math formula*/
.MathJax_Display {
margin: 0em 0em;
}
/* Wider container for better content space */
@media (min-width: 1200px) {
.container { max-width: 1280px; }
}
/* Column layout */
#nav, #content, #toc-sidebar { box-sizing: border-box; }
#content {
padding-left: 15px;
padding-right: 15px;
overflow-wrap: break-word;
word-wrap: break-word;
overflow: hidden;
}
/* On mobile: hide sidebar, show hamburger, mobile overlay */
@media (max-width: 767px) {
#nav { display: none; }
#mobile-menu-btn { display: block !important; }
.site-header h3 { font-size: 16px !important; }
.site-header h3 small { display: none !important; }
.site-header form { display: none; }
}
@media (max-width: 400px) {
.site-header h3 { font-size: 14px !important; }
}
/* Mobile sidebar overlay */
.mobile-sidebar-overlay {
display: none;
position: fixed;
top: 0; left: 0; right: 0; bottom: 0;
z-index: 9999;
background: rgba(0,0,0,0.4);
}
.mobile-sidebar-overlay.open { display: block; }
.mobile-sidebar-panel {
position: fixed;
top: 0; left: 0; bottom: 0;
width: 280px;
background: #fff;
z-index: 10000;
overflow-y: auto;
padding: 16px;
box-shadow: 2px 0 16px rgba(0,0,0,0.15);
transform: translateX(-100%);
transition: transform 0.25s ease;
}
.mobile-sidebar-overlay.open .mobile-sidebar-panel { transform: translateX(0); }
.mobile-sidebar-close {
position: absolute; top: 10px; right: 12px;
background: none; border: none; font-size: 22px; cursor: pointer; color: #666;
}
html.dark .mobile-sidebar-panel { background: #1e293b; }
html.dark .mobile-sidebar-overlay { background: rgba(0,0,0,0.6); }
html.dark .mobile-sidebar-close { color: #94a3b8; }
</style>
<!-- Open Graph -->
<meta property="og:title" content="Algorithmic Fairness in R: fairml & aif360 for Bias Auditing">
<meta property="og:description" content="Audit machine learning models for fairness using R. Learn disparate impact, calibration, and practical bias auditing with fairml and aif360.">
<meta property="og:type" content="article">
<meta property="og:url" content="https://r-statistics.co/Algorithmic-Fairness-in-R.html">
<meta property="og:site_name" content="r-statistics.co">
<meta property="og:image" content="https://r-statistics.co/screenshots/og/Algorithmic-Fairness-in-R.png">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:title" content="Algorithmic Fairness in R: fairml & aif360 for Bias Auditing">
<meta name="twitter:description" content="Audit machine learning models for fairness using R. Learn disparate impact, calibration, and practical bias auditing with fairml and aif360.">
<meta name="twitter:image" content="https://r-statistics.co/screenshots/og/Algorithmic-Fairness-in-R.png">
<!-- JSON-LD Structured Data -->
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": ["TechArticle", "LearningResource"],
"headline": "Algorithmic Fairness in R: fairml & aif360 for Bias Auditing",
"description": "Audit machine learning models for fairness using R. Learn disparate impact, calibration, and practical bias auditing with fairml and aif360.",
"author": {"@type": "Person", "name": "Selva Prabhakaran", "url": "https://r-statistics.co/about/", "jobTitle": "Data Scientist"},
"publisher": {"@type": "Organization", "name": "r-statistics.co", "url": "https://r-statistics.co/", "logo": {"@type": "ImageObject", "url": "https://r-statistics.co/screenshots/og-default.png"}},
"url": "https://r-statistics.co/Algorithmic-Fairness-in-R.html",
"datePublished": "2026-03-29",
"dateModified": "2026-03-29",
"inLanguage": "en",
"educationalLevel": "Intermediate",
"programmingLanguage": "R",
"speakable": {"@type": "SpeakableSpecification", "cssSelector": [".lead", "#content h1"]}
}
</script>
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "BreadcrumbList",
"itemListElement": [
{"@type": "ListItem", "position": 1, "name": "Home", "item": "https://r-statistics.co/"},
{"@type": "ListItem", "position": 2, "name": "Algorithmic Fairness in R: fairml & aif360 for Bias Auditing", "item": "https://r-statistics.co/Algorithmic-Fairness-in-R.html"}
]
}
</script>
<!-- WebR Interactive R Code -->
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/codemirror/5.65.16/codemirror.min.css">
<link href="https://fonts.googleapis.com/css2?family=JetBrains+Mono:wght@400;500;700&display=swap" rel="stylesheet">
<style>
/* --- Loading banner: compact, collapses when ready --- */
.webr-loading-banner {
padding: 8px 16px;
background: #f0f4ff;
color: #3b5998;
text-align: center;
margin: 0 0 8px 0;
border-radius: 6px;
font-size: 13px;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
border: 1px solid #c7d2fe;
transition: all 0.4s ease;
overflow: hidden;
max-height: 50px;
}
.webr-loading-banner.ready {
background: #ecfdf5;
color: #166534;
border-color: #86efac;
}
.webr-loading-banner.hidden {
max-height: 0;
padding: 0 16px;
margin: 0;
border-color: transparent;
opacity: 0;
}
/* --- Run All bar (subtle with background) --- */
.webr-runall-bar {
display: flex;
justify-content: flex-end;
align-items: center;
gap: 10px;
padding: 7px 14px;
margin: 0 0 18px 0;
background: #f8fafb;
border: 1px solid #edf0f3;
border-radius: 8px;
opacity: 0;
transform: translateY(-4px);
transition: opacity 0.3s ease, transform 0.3s ease;
}
.webr-runall-bar.visible { opacity: 1; transform: translateY(0); }
.webr-runall-label {
font-size: 11px;
color: #9ca3af;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
display: flex;
align-items: center;
gap: 5px;
}
.webr-runall-label::before {
content: '';
width: 6px; height: 6px;
border-radius: 50%;
background: #2da44e;
flex-shrink: 0;
}
.webr-runall-btn {
background: none;
color: #3f73d8;
border: 1px solid #d0d7de;
padding: 4px 12px;
border-radius: 6px;
font-size: 11px;
font-weight: 400;
cursor: pointer;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
transition: all 0.15s ease;
}
.webr-runall-btn:hover { background: #f6f8fa; border-color: #3f73d8; }
.webr-runall-btn:disabled { color: #a5b4d4; cursor: wait; }
.webr-runall-btn.running { color: #d29922; border-color: #d29922; animation: pulse-run 1.5s ease-in-out infinite; }
/* --- Code block container --- */
.webr-container {
margin: 22px 0;
border: 1px solid #e1e4e8;
border-radius: 8px;
overflow: hidden;
box-shadow: 0 2px 8px rgba(0,0,0,0.06), 0 0 1px rgba(0,0,0,0.1);
transition: box-shadow 0.2s ease;
}
.webr-container:hover {
box-shadow: 0 4px 16px rgba(0,0,0,0.1), 0 0 1px rgba(0,0,0,0.12);
border-color: #c7d2fe;
}
.webr-container:focus-within {
border-color: #3b82f6;
box-shadow: 0 0 0 3px rgba(59,130,246,0.15), 0 2px 8px rgba(0,0,0,0.06);
}
/* --- Code block header bar --- */
.webr-header {
display: flex;
align-items: center;
justify-content: space-between;
padding: 5px 12px;
background: #f1f5f9;
border-bottom: 1px solid #e2e8f0;
font-family: 'IBM Plex Sans', -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
}
.webr-header-label {
font-size: 11px; font-weight: 600; color: #64748b;
text-transform: uppercase; letter-spacing: 0.5px;
}
.webr-header-badge {
background: #3b82f6; color: #fff; padding: 1px 8px;
border-radius: 4px; font-size: 10px; font-weight: 600; letter-spacing: 0.3px;
}
/* --- Run button shortcut hint --- */
.webr-run-shortcut {
font-size: 10px; opacity: 0.7; margin-left: 6px; font-weight: 400;
}
/* --- Button bar (below code editor) --- */
.webr-buttons {
padding: 6px 10px;
background: #f7f8fa;
border-top: 1px solid #e1e4e8;
display: flex;
justify-content: space-between;
align-items: center;
}
.webr-buttons-left { display: flex; gap: 6px; align-items: center; }
.webr-buttons-right { display: flex; gap: 6px; align-items: center; }
/* --- Keyboard hint --- */
.webr-kbd-hint {
font-size: 10px;
color: #9ca3af;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
}
.webr-kbd-hint kbd {
background: #f0f1f3;
border: 1px solid #d1d5db;
border-radius: 3px;
padding: 1px 4px;
font-size: 10px;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
box-shadow: 0 1px 0 rgba(0,0,0,0.06);
}
/* --- Copy button --- */
.webr-copy-btn {
background: none;
border: 1px solid #d0d7de;
color: #656d76;
padding: 3px 7px;
border-radius: 6px;
font-size: 14px;
cursor: pointer;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
transition: all 0.15s ease;
display: flex;
align-items: center;
gap: 4px;
line-height: 1;
}
.webr-copy-btn:hover { background: #f6f8fa; color: #24292f; border-color: #c0c8d0; }
.webr-copy-btn.copied { color: #2da44e; border-color: #2da44e; }
/* --- Pre-CodeMirror: reserve space to prevent CLS --- */
.webr-editor {
min-height: 80px;
font-family: 'IBM Plex Mono', 'JetBrains Mono', 'Consolas', monospace;
font-size: 14px;
line-height: 1.6;
padding: 8px 12px 8px 40px;
background: #fafbfd;
white-space: pre;
overflow: hidden;
color: #24292f;
}
/* After CodeMirror initializes, it replaces the content — reset padding */
.webr-editor .CodeMirror {
height: auto;
min-height: 48px;
font-size: 14px;
font-family: 'IBM Plex Mono', 'JetBrains Mono', 'Consolas', monospace;
line-height: 1.6;
background: #fafbfd;
padding: 4px 0;
}
.webr-editor.cm-initialized { padding: 0; min-height: auto; }
.webr-editor .CodeMirror-gutters {
background: #f0f2f5;
border-right: 1px solid #e1e4e8;
padding-right: 2px;
}
.webr-editor .CodeMirror-linenumber {
color: #a0a8b4;
font-size: 11px;
padding: 0 6px 0 8px;
min-width: 28px;
}
.webr-editor .CodeMirror-cursor { border-left-color: #3b82f6; border-left-width: 2px; }
.webr-editor .CodeMirror-selected { background: #c8e1ff !important; }
.webr-editor .CodeMirror-activeline-background { background: #f0f4ff; }
.webr-editor .CodeMirror-matchingbracket { color: #0969da !important; font-weight: 700; background: #ddf4ff; border-radius: 2px; }
/* --- R syntax colors (refined) --- */
.webr-editor .cm-keyword { color: #cf222e; font-weight: 500; }
.webr-editor .cm-atom { color: #8250df; }
.webr-editor .cm-number { color: #0550ae; }
.webr-editor .cm-def { color: #8250df; }
.webr-editor .cm-variable { color: #24292f; }
.webr-editor .cm-variable-2 { color: #953800; }
.webr-editor .cm-variable-3 { color: #0550ae; }
.webr-editor .cm-property { color: #0550ae; }
.webr-editor .cm-operator { color: #cf222e; }
.webr-editor .cm-comment { color: #6e7781; font-style: italic; }
.webr-editor .cm-string { color: #0a3069; }
.webr-editor .cm-string-2 { color: #0a3069; }
.webr-editor .cm-builtin { color: #8250df; font-weight: 500; }
/* --- Run button --- */
.webr-run-btn {
background: #3f73d8;
color: #fff;
border: none;
padding: 4px 14px;
border-radius: 6px;
font-size: 12px;
font-weight: 500;
cursor: pointer;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
transition: all 0.15s ease;
}
.webr-run-btn:hover { background: #2c5fbe; }
.webr-run-btn:active { background: #254fa0; }
.webr-run-btn:disabled { background: #a5b4d4; cursor: wait; }
.webr-run-btn.running { background: #d29922; animation: pulse-run 1.5s ease-in-out infinite; }
@keyframes pulse-run { 0%, 100% { opacity: 1; } 50% { opacity: 0.8; } }
/* --- Reset button --- */
.webr-reset-btn {
background: none;
color: #656d76;
border: 1px solid #d0d7de;
padding: 3px 10px;
border-radius: 6px;
font-size: 12px;
cursor: pointer;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
transition: all 0.15s ease;
}
.webr-reset-btn:hover { background: #f6f8fa; color: #24292f; border-color: #c0c8d0; }
/* --- Output panel --- */
.webr-output {
margin: 0;
padding: 14px 16px;
background: #f1f5f9;
color: #1e293b;
font-family: 'IBM Plex Mono', 'JetBrains Mono', 'Consolas', monospace;
font-size: 13px;
line-height: 1.6;
max-height: 400px;
overflow-y: auto;
white-space: pre-wrap;
display: none;
border-top: 1px solid #e2e8f0;
}
.webr-output::-webkit-scrollbar { width: 8px; }
.webr-output::-webkit-scrollbar-track { background: #f1f5f9; }
.webr-output::-webkit-scrollbar-thumb { background: #cbd5e1; border-radius: 4px; }
.webr-output.has-content { display: block; }
.webr-output.has-error { color: #dc2626; border-top-color: #dc2626; }
/* --- Plot output --- */
.webr-plot-output { text-align: center; padding: 16px; background: #fff; display: none; border-top: 1px solid #e1e4e8; }
.webr-plot-output.has-content { display: block; }
.webr-plot-output img, .webr-plot-output canvas { max-width: 100%; height: auto; border-radius: 6px; box-shadow: 0 1px 4px rgba(0,0,0,0.08); }
.mermaid { margin: 15px 0; text-align: center; }
/* ===== DARK MODE — WebR components ===== */
/* Loading banner */
html.dark .webr-loading-banner { background: #1e293b; color: #94a3b8; border-color: #334155; }
html.dark .webr-loading-banner.ready { background: #1e293b; color: #4ade80; border-color: #334155; }
/* Run All bar */
html.dark .webr-runall-bar { background: #1e293b; border-color: #334155; }
html.dark .webr-runall-label { color: #94a3b8; }
html.dark .webr-runall-label::before { background: #4ade80; }
html.dark .webr-runall-btn { color: #94a3b8; border-color: #475569; }
html.dark .webr-runall-btn:hover { background: #334155; color: #e2e8f0; border-color: #3b82f6; }
/* Container */
html.dark .webr-container { border-color: #334155; box-shadow: 0 2px 8px rgba(0,0,0,0.3); }
html.dark .webr-container:hover { border-color: #4f46e5; box-shadow: 0 4px 16px rgba(0,0,0,0.4); }
html.dark .webr-container:focus-within { border-color: #3b82f6; box-shadow: 0 0 0 3px rgba(59,130,246,0.25); }
/* Header bar */
html.dark .webr-header { background: #0f172a; border-bottom-color: #1e293b; }
html.dark .webr-header-label { color: #64748b; }
html.dark .webr-header-badge { background: #3b82f6; color: #fff; }
/* Button bar */
html.dark .webr-buttons { background: #0f172a; border-top-color: #1e293b; }
/* Run button */
html.dark .webr-run-btn { background: #3b82f6; }
html.dark .webr-run-btn:hover { background: #2563eb; }
html.dark .webr-run-shortcut { opacity: 0.6; }
/* Reset button */
html.dark .webr-reset-btn { color: #94a3b8; border-color: #475569; }
html.dark .webr-reset-btn:hover { background: #334155; color: #e2e8f0; }
/* Copy button */
html.dark .webr-copy-btn { color: #94a3b8; border-color: #475569; }
html.dark .webr-copy-btn:hover { background: #334155; color: #e2e8f0; border-color: #64748b; }
html.dark .webr-copy-btn.copied { color: #4ade80; border-color: #4ade80; }
/* CodeMirror editor */
html.dark .webr-editor .CodeMirror { background: #1e293b; color: #e2e8f0; }
html.dark .webr-editor .CodeMirror-gutters { background: #1e293b; border-right-color: #334155; }
html.dark .webr-editor .CodeMirror-linenumber { color: #475569; }
html.dark .webr-editor .CodeMirror-cursor { border-left-color: #60a5fa; }
html.dark .webr-editor .CodeMirror-selected { background: #334155 !important; }
html.dark .webr-editor .CodeMirror-activeline-background { background: #1e3a5f; }
html.dark .webr-editor .CodeMirror-matchingbracket { color: #60a5fa !important; background: #1e3a5f; }
html.dark .webr-editor .cm-keyword { color: #f87171; }
html.dark .webr-editor .cm-atom { color: #c084fc; }
html.dark .webr-editor .cm-number { color: #60a5fa; }
html.dark .webr-editor .cm-def { color: #c084fc; }
html.dark .webr-editor .cm-variable { color: #e2e8f0; }
html.dark .webr-editor .cm-variable-2 { color: #fb923c; }
html.dark .webr-editor .cm-variable-3 { color: #60a5fa; }
html.dark .webr-editor .cm-property { color: #60a5fa; }
html.dark .webr-editor .cm-operator { color: #f87171; }
html.dark .webr-editor .cm-comment { color: #64748b; }
html.dark .webr-editor .cm-string { color: #86efac; }
html.dark .webr-editor .cm-string-2 { color: #86efac; }
html.dark .webr-editor .cm-builtin { color: #c084fc; }
/* Output panel */
html.dark .webr-output { background: #0f172a; color: #cbd5e1; border-top-color: #1e293b; }
html.dark .webr-output::-webkit-scrollbar-track { background: #0f172a; }
html.dark .webr-output::-webkit-scrollbar-thumb { background: #475569; }
html.dark .webr-output.has-error { color: #f87171; border-top-color: #f87171; }
/* Plot output */
html.dark .webr-plot-output { background: #1e293b; border-top-color: #334155; }
/* Kbd hint */
html.dark .webr-kbd-hint { color: #64748b; }
html.dark .webr-kbd-hint kbd { background: #334155; border-color: #475569; color: #94a3b8; }
</style>
<!-- Add Google search -->
<script language="Javascript" type="text/javascript">
function my_search_google()
{
var query = document.getElementById("my-google-search").value;
window.open("https://google.com/search?q=" + query
+ "%20site:" + "https://r-statistics.co");
}
</script>
</head>
<body>
<!-- Mobile sidebar overlay -->
<div class="mobile-sidebar-overlay" id="mobile-sidebar">
<div class="mobile-sidebar-panel">
<button class="mobile-sidebar-close" id="mobile-sidebar-close">×</button>
<div id="mobile-sidebar-content">
<!-- Populated by toc.js -->
</div>
</div>
</div>
<div class="container">
<div class="masthead">
<!--
<ul class="nav nav-pills pull-right">
<li class="dropdown">
<a href="#" class="dropdown-toggle" data-toggle="dropdown">
Table of contents<b class="caret"></b>
</a>
<ul class="dropdown-menu pull-right" role="menu">
<li class="dropdown-header"></li>
<li class="dropdown-header">Tutorial</li>
<li><a href="R-Tutorial.html">R Tutorial</a></li>
<li class="dropdown-header">ggplot2</li>
<li><a href="ggplot2-Tutorial-With-R.html">ggplot2 Short Tutorial</a></li>
<li><a href="Complete-Ggplot2-Tutorial-Part1-With-R-Code.html">ggplot2 Tutorial 1 - Intro</a></li>
<li><a href="Complete-Ggplot2-Tutorial-Part2-Customizing-Theme-With-R-Code.html">ggplot2 Tutorial 2 - Theme</a></li>
<li><a href="Top50-Ggplot2-Visualizations-MasterList-R-Code.html">ggplot2 Tutorial 3 - Masterlist</a></li>
<li><a href="ggplot2-cheatsheet.html">ggplot2 Quickref</a></li>
<li class="dropdown-header">Foundations</li>
<li><a href="Linear-Regression.html">Linear Regression</a></li>
<li><a href="Statistical-Tests-in-R.html">Statistical Tests</a></li>
<li><a href="Missing-Value-Treatment-With-R.html">Missing Value Treatment</a></li>
<li><a href="Outlier-Treatment-With-R.html">Outlier Analysis</a></li>
<li><a href="Variable-Selection-and-Importance-With-R.html">Feature Selection</a></li>
<li><a href="Model-Selection-in-R.html">Model Selection</a></li>
<li><a href="Logistic-Regression-With-R.html">Logistic Regression</a></li>
<li><a href="Environments.html">Advanced Linear Regression</a></li>
<li class="dropdown-header">Advanced Regression Models</li>
<li><a href="adv-regression-models.html">Advanced Regression Models</a></li>
<li class="dropdown-header">Time Series</li>
<li><a href="Time-Series-Analysis-With-R.html">Time Series Analysis</a></li>
<li><a href="Time-Series-Forecasting-With-R.html">Time Series Forecasting </a></li>
<li><a href="Time-Series-Forecasting-With-R-part2.html">More Time Series Forecasting</a></li>
<li class="dropdown-header">High Performance Computing</li>
<li><a href="Parallel-Computing-With-R.html">Parallel computing</a></li>
<li><a href="Strategies-To-Improve-And-Speedup-R-Code.html">Strategies to Speedup R code</a></li>
<li class="dropdown-header">Useful Techniques</li>
<li><a href="Association-Mining-With-R.html">Association Mining</a></li>
<li><a href="Multi-Dimensional-Scaling-With-R.html">Multi Dimensional Scaling</a></li>
<li><a href="Profiling.html">Optimization</a></li>
<li><a href="Information-Value-With-R.html">InformationValue package</a></li>
</ul>
</li>
</ul>
-->
<div class="site-header" style="display:flex; justify-content:space-between; align-items:center; padding: 12px 0 8px 0; flex-wrap: nowrap; gap: 10px;">
<div style="display:flex; align-items:center; gap:8px; min-width:0;">
<button id="mobile-menu-btn" aria-label="Menu" style="display:none;background:none;border:1px solid #d0d7de;border-radius:6px;padding:4px 8px;cursor:pointer;font-size:18px;line-height:1;color:#666;">☰</button>
<h3 style="margin:0;white-space:nowrap;"><a href="/" style="text-decoration:none;">r‑statistics.co</a><small style="color:#999; margin-left:4px; font-size:11px;"> by Selva Prabhakaran</small></h3>
</div>
<div style="display:flex; align-items:center; gap:12px; flex-shrink:0;">
<a href="/about/" style="font-size:13px; color:#666; text-decoration:none;">About</a>
<button id="dark-mode-toggle" aria-label="Toggle dark mode" title="Toggle dark mode"
style="background:none;border:1px solid #d0d7de;border-radius:6px;padding:4px 8px;cursor:pointer;font-size:16px;line-height:1;color:#666;transition:all 0.2s;">
☽
</button>
<form onsubmit="my_search_google(); return false;" style="margin:0;">
<input type="text" class="form-control" id="my-google-search" placeholder="Search.." style="width:130px; height:30px; font-size:12px; border-radius:6px;">
</form>
</div>
</div>
<hr style="margin:0 0 16px 0;">
</div>
<div class="row">
<div class="col-xs-12 col-sm-3" id="nav">
<div id="sidebar-nav">
<!-- Loaded dynamically from www/sidebar.json by toc.js -->
</div>
</div>
<div id="content" class="col-xs-12 col-sm-7">
<nav class="breadcrumb-nav" aria-label="Breadcrumb"><a href="/">Home</a> <span class="breadcrumb-sep">›</span> <span class="breadcrumb-current">Algorithmic Fairness in R: fairml & aif360 for Bias Auditing</span></nav>
<h1>Algorithmic Fairness in R: fairml & aif360 for Bias Auditing</h1>
<p class="lead">Algorithmic fairness ensures that machine learning models don't systematically discriminate against protected groups. This guide teaches you to measure, audit, and improve fairness using R tools — because a model that's accurate on average can still be unfair to specific groups.</p>
<p>A hiring model that rejects 80% of female applicants but only 30% of male applicants is unfair — even if its overall accuracy is high. A credit scoring model that gives higher rates to minorities with the same creditworthiness as non-minorities is unfair. These aren't hypothetical: they've happened in real deployments. This guide gives you the tools to catch and fix these problems.</p>
<h2>Fairness Definitions</h2>
<p>There are multiple definitions of fairness, and they can conflict with each other. Understanding them is essential for choosing the right one for your context.</p>
<table class="table table-striped">
<thead>
<tr>
<th>Definition</th>
<th>Meaning</th>
<th>Formula</th>
</tr>
</thead>
<tbody>
<tr>
<td>Demographic parity</td>
<td>Equal selection rates across groups</td>
<td>P(Y=1\</td>
<td>A=a) = P(Y=1\</td>
<td>A=b)</td>
</tr>
<tr>
<td>Equalized odds</td>
<td>Equal TPR and FPR across groups</td>
<td>P(Yhat=1\</td>
<td>Y=y,A=a) = P(Yhat=1\</td>
<td>Y=y,A=b)</td>
</tr>
<tr>
<td>Equal opportunity</td>
<td>Equal TPR across groups</td>
<td>P(Yhat=1\</td>
<td>Y=1,A=a) = P(Yhat=1\</td>
<td>Y=1,A=b)</td>
</tr>
<tr>
<td>Calibration</td>
<td>Same meaning of scores across groups</td>
<td>P(Y=1\</td>
<td>Score=s,A=a) = P(Y=1\</td>
<td>Score=s,A=b)</td>
</tr>
<tr>
<td>Predictive parity</td>
<td>Equal precision across groups</td>
<td>P(Y=1\</td>
<td>Yhat=1,A=a) = P(Y=1\</td>
<td>Yhat=1,A=b)</td>
</tr>
<tr>
<td>Individual fairness</td>
<td>Similar individuals treated similarly</td>
<td>d(x,x') small implies d(f(x),f(x')) small</td>
</tr>
</tbody>
</table>
<h4>The Impossibility Theorem</h4>
<p>A critical result: except in trivial cases, you <strong>cannot</strong> simultaneously satisfy demographic parity, equalized odds, and calibration. You must choose which fairness criterion matters most for your application.</p>
<div class="webr-container">
<div class="webr-code-block">
<div class="webr-editor" data-language="r"># Demonstrating the fairness trade-off
set.seed(42)
n <- 1000
# Simulate two groups with different base rates
group <- rep(c("A","B"), each = n/2)
base_rate <- ifelse(group == "A", 0.4, 0.2) # Different base rates
true_label <- rbinom(n, 1, base_rate)
# A perfectly calibrated model
score <- true_label + rnorm(n, 0, 0.3)
predicted <- as.integer(score > 0.5)
cat("=== Fairness Trade-off Demo ===\n")
cat("Base rates differ between groups:\n")
cat(" Group A base rate:", mean(true_label[group == "A"]), "\n")
cat(" Group B base rate:", mean(true_label[group == "B"]), "\n")
cat("\nSelection rates (demographic parity check):\n")
cat(" Group A:", mean(predicted[group == "A"]), "\n")
cat(" Group B:", mean(predicted[group == "B"]), "\n")
cat("\nTrue Positive Rates (equal opportunity check):\n")
tpr_a <- mean(predicted[group == "A" & true_label == 1])
tpr_b <- mean(predicted[group == "B" & true_label == 1])
cat(" Group A TPR:", round(tpr_a, 3), "\n")
cat(" Group B TPR:", round(tpr_b, 3), "\n")
cat("\nEqualizing selection rates would break calibration.\n")
cat("Equalizing TPR would change selection rates.\n")
cat("You must choose which fairness criterion to prioritize.\n")</div>
<div class="webr-buttons">
<button class="btn btn-sm btn-primary webr-run-btn" onclick="runWebR(this)">▶ Run</button>
<button class="btn btn-sm btn-default webr-reset-btn" onclick="resetWebR(this)">↺ Reset</button>
</div>
<pre class="webr-output"></pre>
</div>
</div>
<h2>Measuring Disparate Impact</h2>
<p>The four-fifths rule: the selection rate for any protected group should be at least 80% of the rate for the most-selected group.</p>
<div class="webr-container">
<div class="webr-code-block">
<div class="webr-editor" data-language="r"># Comprehensive disparate impact analysis
set.seed(42)
n <- 800
applicants <- data.frame(
gender = sample(c("Male","Female"), n, replace = TRUE),
score = rnorm(n, 70, 10)
)
# Biased threshold: unconsciously favoring one group
applicants$hired <- with(applicants, {
threshold <- ifelse(gender == "Male", 65, 70)
as.integer(score > threshold)
})
# Disparate impact analysis
cat("=== Disparate Impact Analysis ===\n")
hire_rates <- tapply(applicants$hired, applicants$gender, mean)
cat("Hiring rates:\n")
print(round(hire_rates, 3))
di_ratio <- min(hire_rates) / max(hire_rates)
cat(sprintf("\nDisparate impact ratio: %.3f\n", di_ratio))
cat(sprintf("Four-fifths threshold: 0.800\n"))
cat(sprintf("Result: %s\n", ifelse(di_ratio >= 0.8, "PASS", "FAIL - potential bias")))
# Statistical significance
cat("\n=== Chi-square test ===\n")
ct <- table(applicants$gender, applicants$hired)
chi <- chisq.test(ct)
cat(sprintf("Chi-squared = %.2f, p = %.4f\n", chi$statistic, chi$p.value))</div>
<div class="webr-buttons">
<button class="btn btn-sm btn-primary webr-run-btn" onclick="runWebR(this)">▶ Run</button>
<button class="btn btn-sm btn-default webr-reset-btn" onclick="resetWebR(this)">↺ Reset</button>
</div>
<pre class="webr-output"></pre>
</div>
</div>
<h2>Building a Fairness Audit Function</h2>
<div class="webr-container">
<div class="webr-code-block">
<div class="webr-editor" data-language="r"># Reusable fairness audit function
fairness_audit <- function(actual, predicted, group, positive_label = 1) {
groups <- unique(group)
results <- data.frame(
Group = character(),
N = integer(),
SelectionRate = numeric(),
TPR = numeric(),
FPR = numeric(),
Precision = numeric(),
stringsAsFactors = FALSE
)
for (g in groups) {
mask <- group == g
a <- actual[mask]
p <- predicted[mask]
tp <- sum(a == positive_label & p == positive_label)
fp <- sum(a != positive_label & p == positive_label)
fn <- sum(a == positive_label & p != positive_label)
tn <- sum(a != positive_label & p != positive_label)
results <- rbind(results, data.frame(
Group = g,
N = sum(mask),
SelectionRate = round(mean(p == positive_label), 3),
TPR = round(ifelse(tp+fn > 0, tp/(tp+fn), NA), 3),
FPR = round(ifelse(fp+tn > 0, fp/(fp+tn), NA), 3),
Precision = round(ifelse(tp+fp > 0, tp/(tp+fp), NA), 3),
stringsAsFactors = FALSE
))
}
# Calculate disparities
cat("=== Fairness Audit Report ===\n\n")
print(results, row.names = FALSE)
cat("\n--- Disparity Ratios ---\n")
max_sr <- max(results$SelectionRate)
for (i in 1:nrow(results)) {
ratio <- results$SelectionRate[i] / max_sr
status <- ifelse(ratio >= 0.8, "OK", "WARNING")
cat(sprintf(" %s: SR ratio = %.3f [%s]\n",
results$Group[i], ratio, status))
}
cat(sprintf("\n--- Equal Opportunity (TPR) ---\n"))
max_tpr <- max(results$TPR, na.rm = TRUE)
for (i in 1:nrow(results)) {
ratio <- results$TPR[i] / max_tpr
cat(sprintf(" %s: TPR ratio = %.3f\n", results$Group[i], ratio))
}
}
# Test the audit function
set.seed(42)
n <- 600
test_data <- data.frame(
group = sample(c("Young","Middle","Senior"), n, replace = TRUE, prob = c(0.4,0.4,0.2)),
actual = rbinom(n, 1, 0.3),
predicted = rbinom(n, 1, 0.28)
)
fairness_audit(test_data$actual, test_data$predicted, test_data$group)</div>
<div class="webr-buttons">
<button class="btn btn-sm btn-primary webr-run-btn" onclick="runWebR(this)">▶ Run</button>
<button class="btn btn-sm btn-default webr-reset-btn" onclick="resetWebR(this)">↺ Reset</button>
</div>
<pre class="webr-output"></pre>
</div>
</div>
<h2>Fairness Packages in R</h2>
<h4>fairml Package</h4>
<p>The <code>fairml</code> package implements fair regression and classification models that explicitly include fairness constraints.</p>
<div class="webr-container">
<div class="webr-code-block">
<div class="webr-editor" data-language="r"># fairml concepts (demonstration without package dependency)
cat("=== fairml Package Overview ===\n\n")
cat("Key functions:\n")
cat(" frrm() - Fair Ridge Regression Model\n")
cat(" fgrrm() - Fair Generalized Ridge Regression\n")
cat(" nclm() - Nonconvex Penalized Logistic Model\n")
cat(" zlrm() - Zafar Logistic Regression Model\n\n")
cat("Usage pattern:\n")
cat(' library(fairml)\n')
cat(' model <- frrm(y ~ x1 + x2, data = df,\n')
cat(' sensitive = df$protected_attr,\n')
cat(' unfairness = 0.05) # Max allowed unfairness\n\n')
cat("The unfairness parameter (0 to 1) controls the trade-off:\n")
cat(" 0.00 = Perfectly fair (may sacrifice accuracy)\n")
cat(" 0.05 = 5% unfairness tolerance (good balance)\n")
cat(" 1.00 = No fairness constraint (standard model)\n")</div>
<div class="webr-buttons">
<button class="btn btn-sm btn-primary webr-run-btn" onclick="runWebR(this)">▶ Run</button>
<button class="btn btn-sm btn-default webr-reset-btn" onclick="resetWebR(this)">↺ Reset</button>
</div>
<pre class="webr-output"></pre>
</div>
</div>
<h4>aif360 (AI Fairness 360)</h4>
<div class="webr-container">
<div class="webr-code-block">
<div class="webr-editor" data-language="r">cat("=== AIF360 for R ===\n\n")
cat("IBM's AI Fairness 360 toolkit (Python-based, R interface available):\n\n")
cat("Bias metrics:\n")
cat(" - Statistical parity difference\n")
cat(" - Disparate impact ratio\n")
cat(" - Equal opportunity difference\n")
cat(" - Average odds difference\n")
cat(" - Theil index\n\n")
cat("Bias mitigation algorithms:\n")
cat(" Pre-processing: Reweighting, Optimized Preprocessing\n")
cat(" In-processing: Adversarial Debiasing, Prejudice Remover\n")
cat(" Post-processing: Equalized Odds, Calibrated Equalized Odds\n\n")
cat("R usage via reticulate:\n")
cat(' library(reticulate)\n')
cat(' aif <- import("aif360.datasets")\n')
cat(' metrics <- import("aif360.metrics")\n')</div>
<div class="webr-buttons">
<button class="btn btn-sm btn-primary webr-run-btn" onclick="runWebR(this)">▶ Run</button>
<button class="btn btn-sm btn-default webr-reset-btn" onclick="resetWebR(this)">↺ Reset</button>
</div>
<pre class="webr-output"></pre>
</div>
</div>
<h2>Practical Audit Workflow</h2>
<table class="table table-striped">
<thead>
<tr>
<th>Step</th>
<th>Action</th>
<th>Tool</th>
</tr>
</thead>
<tbody>
<tr>
<td>1. Define protected attributes</td>
<td>List sensitive variables</td>
<td>Domain knowledge</td>
</tr>
<tr>
<td>2. Choose fairness metric</td>
<td>Match to application context</td>
<td>See definitions table above</td>
</tr>
<tr>
<td>3. Measure baseline</td>
<td>Calculate metrics on current model</td>
<td><code>fairness_audit()</code> function</td>
</tr>
<tr>
<td>4. Set threshold</td>
<td>Define acceptable disparity level</td>
<td>Four-fifths rule or domain-specific</td>
</tr>
<tr>
<td>5. Mitigate if needed</td>
<td>Apply debiasing technique</td>
<td>fairml, reweighting, threshold tuning</td>
</tr>
<tr>
<td>6. Re-measure</td>
<td>Verify improvement</td>
<td>Same audit function</td>
</tr>
<tr>
<td>7. Document</td>
<td>Record decisions and trade-offs</td>
<td>Analysis report</td>
</tr>