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
|
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=US-ASCII">
<title>Geometric Distribution Examples</title>
<link rel="stylesheet" href="../../../../../../../../../doc/src/boostbook.css" type="text/css">
<meta name="generator" content="DocBook XSL Stylesheets V1.76.1">
<link rel="home" href="../../../../index.html" title="Math Toolkit">
<link rel="up" href="../weg.html" title="Worked Examples">
<link rel="prev" href="binom_eg/binom_size_eg.html" title="Estimating Sample Sizes for a Binomial Distribution.">
<link rel="next" href="neg_binom_eg.html" title="Negative Binomial Distribution Examples">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
<table cellpadding="2" width="100%"><tr>
<td valign="top"><img alt="Boost C++ Libraries" width="277" height="86" src="../../../../../../../../../boost.png"></td>
<td align="center"><a href="../../../../../../../../../index.html">Home</a></td>
<td align="center"><a href="../../../../../../../../../libs/libraries.htm">Libraries</a></td>
<td align="center"><a href="http://www.boost.org/users/people.html">People</a></td>
<td align="center"><a href="http://www.boost.org/users/faq.html">FAQ</a></td>
<td align="center"><a href="../../../../../../../../../more/index.htm">More</a></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="binom_eg/binom_size_eg.html"><img src="../../../../../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../weg.html"><img src="../../../../../../../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../../../../index.html"><img src="../../../../../../../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="neg_binom_eg.html"><img src="../../../../../../../../../doc/src/images/next.png" alt="Next"></a>
</div>
<div class="section math_toolkit_dist_stat_tut_weg_geometric_eg">
<div class="titlepage"><div><div><h5 class="title">
<a name="math_toolkit.dist.stat_tut.weg.geometric_eg"></a><a class="link" href="geometric_eg.html" title="Geometric Distribution Examples">Geometric
Distribution Examples</a>
</h5></div></div></div>
<p>
For this example, we will opt to #define two macros to control the error
and discrete handling policies. For this simple example, we want to avoid
throwing an exception (the default policy) and just return infinity.
We want to treat the distribution as if it was continuous, so we choose
a discrete_quantile policy of real, rather than the default policy integer_round_outwards.
</p>
<p>
</p>
<pre class="programlisting"><span class="preprocessor">#define</span> <span class="identifier">BOOST_MATH_OVERFLOW_ERROR_POLICY</span> <span class="identifier">ignore_error</span>
<span class="preprocessor">#define</span> <span class="identifier">BOOST_MATH_DISCRETE_QUANTILE_POLICY</span> <span class="identifier">real</span>
</pre>
<p>
</p>
<div class="caution"><table border="0" summary="Caution">
<tr>
<td rowspan="2" align="center" valign="top" width="25"><img alt="[Caution]" src="../../../../../../../../../doc/src/images/caution.png"></td>
<th align="left">Caution</th>
</tr>
<tr><td align="left" valign="top"><p>
It is vital to #include distributions etc <span class="bold"><strong>after</strong></span>
the above #defines
</p></td></tr>
</table></div>
<p>
After that we need some includes to provide easy access to the negative
binomial distribution, and we need some std library iostream, of course.
</p>
<p>
</p>
<pre class="programlisting"><span class="preprocessor">#include</span> <span class="special"><</span><span class="identifier">boost</span><span class="special">/</span><span class="identifier">math</span><span class="special">/</span><span class="identifier">distributions</span><span class="special">/</span><span class="identifier">geometric</span><span class="special">.</span><span class="identifier">hpp</span><span class="special">></span>
<span class="comment">// for geometric_distribution</span>
<span class="keyword">using</span> <span class="special">::</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">geometric_distribution</span><span class="special">;</span> <span class="comment">// </span>
<span class="keyword">using</span> <span class="special">::</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">geometric</span><span class="special">;</span> <span class="comment">// typedef provides default type is double.</span>
<span class="keyword">using</span> <span class="special">::</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">pdf</span><span class="special">;</span> <span class="comment">// Probability mass function.</span>
<span class="keyword">using</span> <span class="special">::</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">cdf</span><span class="special">;</span> <span class="comment">// Cumulative density function.</span>
<span class="keyword">using</span> <span class="special">::</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">quantile</span><span class="special">;</span>
<span class="preprocessor">#include</span> <span class="special"><</span><span class="identifier">boost</span><span class="special">/</span><span class="identifier">math</span><span class="special">/</span><span class="identifier">distributions</span><span class="special">/</span><span class="identifier">negative_binomial</span><span class="special">.</span><span class="identifier">hpp</span><span class="special">></span>
<span class="comment">// for negative_binomial_distribution</span>
<span class="keyword">using</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">negative_binomial</span><span class="special">;</span> <span class="comment">// typedef provides default type is double.</span>
<span class="preprocessor">#include</span> <span class="special"><</span><span class="identifier">boost</span><span class="special">/</span><span class="identifier">math</span><span class="special">/</span><span class="identifier">distributions</span><span class="special">/</span><span class="identifier">normal</span><span class="special">.</span><span class="identifier">hpp</span><span class="special">></span>
<span class="comment">// for negative_binomial_distribution</span>
<span class="keyword">using</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">normal</span><span class="special">;</span> <span class="comment">// typedef provides default type is double.</span>
<span class="preprocessor">#include</span> <span class="special"><</span><span class="identifier">iostream</span><span class="special">></span>
<span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">cout</span><span class="special">;</span> <span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">endl</span><span class="special">;</span>
<span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">noshowpoint</span><span class="special">;</span> <span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">fixed</span><span class="special">;</span> <span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">right</span><span class="special">;</span> <span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">left</span><span class="special">;</span>
<span class="preprocessor">#include</span> <span class="special"><</span><span class="identifier">iomanip</span><span class="special">></span>
<span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">setprecision</span><span class="special">;</span> <span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">setw</span><span class="special">;</span>
<span class="preprocessor">#include</span> <span class="special"><</span><span class="identifier">limits</span><span class="special">></span>
<span class="keyword">using</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">numeric_limits</span><span class="special">;</span>
</pre>
<p>
</p>
<p>
It is always sensible to use try and catch blocks because defaults policies
are to throw an exception if anything goes wrong.
</p>
<p>
Simple try'n'catch blocks (see below) will ensure that you get a helpful
error message instead of an abrupt (and silent) program abort.
</p>
<h6>
<a name="math_toolkit.dist.stat_tut.weg.geometric_eg.h0"></a>
<span><a name="math_toolkit.dist.stat_tut.weg.geometric_eg.throwing_a_dice"></a></span><a class="link" href="geometric_eg.html#math_toolkit.dist.stat_tut.weg.geometric_eg.throwing_a_dice">Throwing
a dice</a>
</h6>
<p>
The Geometric distribution describes the probability (<span class="emphasis"><em>p</em></span>)
of a number of failures to get the first success in <span class="emphasis"><em>k</em></span>
Bernoulli trials. (A <a href="http://en.wikipedia.org/wiki/Bernoulli_distribution" target="_top">Bernoulli
trial</a> is one with only two possible outcomes, success of failure,
and <span class="emphasis"><em>p</em></span> is the probability of success).
</p>
<p>
Suppose an 'fair' 6-face dice is thrown repeatedly:
</p>
<p>
</p>
<pre class="programlisting"><span class="keyword">double</span> <span class="identifier">success_fraction</span> <span class="special">=</span> <span class="number">1.</span><span class="special">/</span><span class="number">6</span><span class="special">;</span> <span class="comment">// success_fraction (p) = 0.1666</span>
<span class="comment">// (so failure_fraction is 1 - success_fraction = 5./6 = 1- 0.1666 = 0.8333)</span>
</pre>
<p>
</p>
<p>
If the dice is thrown repeatedly until the <span class="bold"><strong>first</strong></span>
time a <span class="emphasis"><em>three</em></span> appears. The probablility distribution
of the number of times it is thrown <span class="bold"><strong>not</strong></span>
getting a <span class="emphasis"><em>three</em></span> (<span class="emphasis"><em>not-a-threes</em></span>
number of failures to get a <span class="emphasis"><em>three</em></span>) is a geometric
distribution with the success_fraction = 1/6 = 0.1666 ̇.
</p>
<p>
We therefore start by constructing a geometric distribution with the
one parameter success_fraction, the probability of success.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">geometric</span> <span class="identifier">g6</span><span class="special">(</span><span class="identifier">success_fraction</span><span class="special">);</span> <span class="comment">// type double by default.</span>
</pre>
<p>
</p>
<p>
To confirm, we can echo the success_fraction parameter of the distribution.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"success fraction of a six-sided dice is "</span> <span class="special"><<</span> <span class="identifier">g6</span><span class="special">.</span><span class="identifier">success_fraction</span><span class="special">()</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span>
</pre>
<p>
</p>
<p>
So the probability of getting a three at the first throw (zero failures)
is
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">0</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.1667</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">cdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">0</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.1667</span>
</pre>
<p>
</p>
<p>
Note that the cdf and pdf are identical because the is only one throw.
If we want the probability of getting the first <span class="emphasis"><em>three</em></span>
on the 2nd throw:
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">1</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.1389</span>
</pre>
<p>
</p>
<p>
If we want the probability of getting the first <span class="emphasis"><em>three</em></span>
on the 1st or 2nd throw (allowing one failure):
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"pdf(g6, 0) + pdf(g6, 1) = "</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">0</span><span class="special">)</span> <span class="special">+</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">1</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span>
</pre>
<p>
</p>
<p>
Or more conveniently, and more generally, we can use the Cumulative Distribution
Function CDF.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"cdf(g6, 1) = "</span> <span class="special"><<</span> <span class="identifier">cdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">1</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.3056</span>
</pre>
<p>
</p>
<p>
If we allow many more (12) throws, the probability of getting our <span class="emphasis"><em>three</em></span>
gets very high:
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"cdf(g6, 12) = "</span> <span class="special"><<</span> <span class="identifier">cdf</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">12</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.9065 or 90% probability.</span>
</pre>
<p>
</p>
<p>
If we want to be much more confident, say 99%, we can estimate the number
of throws to be this sure using the inverse or quantile.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"quantile(g6, 0.99) = "</span> <span class="special"><<</span> <span class="identifier">quantile</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">0.99</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 24.26</span>
</pre>
<p>
</p>
<p>
Note that the value returned is not an integer: if you want an integer
result you should use either floor, round or ceil functions, or use the
policies mechanism. See <a class="link" href="../../../policy/pol_tutorial/understand_dis_quant.html" title="Understanding Quantiles of Discrete Distributions">Understanding
Quantiles of Discrete Distributions</a>
</p>
<p>
The geometric distribution is related to the negative binomial    <code class="computeroutput"><span class="identifier">negative_binomial_distribution</span><span class="special">(</span><span class="identifier">RealType</span> <span class="identifier">r</span><span class="special">,</span> <span class="identifier">RealType</span>
<span class="identifier">p</span><span class="special">);</span></code>
with parameter <span class="emphasis"><em>r</em></span> = 1. So we could get the same result
using the negative binomial, but using the geometric the results will
be faster, and may be more accurate.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">negative_binomial</span> <span class="identifier">nb</span><span class="special">(</span><span class="number">1</span><span class="special">,</span> <span class="identifier">success_fraction</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">nb</span><span class="special">,</span> <span class="number">1</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.1389</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">cdf</span><span class="special">(</span><span class="identifier">nb</span><span class="special">,</span> <span class="number">1</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.3056</span>
</pre>
<p>
</p>
<p>
We could also the complement to express the required probability as 1
- 0.99 = 0.01 (and get the same result):
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"quantile(complement(g6, 1 - p)) "</span> <span class="special"><<</span> <span class="identifier">quantile</span><span class="special">(</span><span class="identifier">complement</span><span class="special">(</span><span class="identifier">g6</span><span class="special">,</span> <span class="number">0.01</span><span class="special">))</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 24.26</span>
</pre>
<p>
</p>
<p>
Note too that Boost.Math geometric distribution is implemented as a continuous
function. Unlike other implementations (for example R) it <span class="bold"><strong>uses</strong></span>
the number of failures as a <span class="bold"><strong>real</strong></span> parameter,
not as an integer. If you want this integer behaviour, you may need to
enforce this by rounding the parameter you pass, probably rounding down,
to the nearest integer. For example, R returns the success fraction probability
for all values of failures from 0 to 0.999999 thus:
</p>
<pre class="programlisting">   R> formatC(pgeom(0.0001,0.5, FALSE), digits=17) " 0.5"
</pre>
<p>
So in Boost.Math the equivalent is
</p>
<p>
</p>
<pre class="programlisting"> <span class="identifier">geometric</span> <span class="identifier">g05</span><span class="special">(</span><span class="number">0.5</span><span class="special">);</span> <span class="comment">// Probability of success = 0.5 or 50%</span>
<span class="comment">// Output all potentially significant digits for the type, here double.</span>
<span class="preprocessor">#ifdef</span> <span class="identifier">BOOST_NO_NUMERIC_LIMITS_LOWEST</span>
<span class="keyword">int</span> <span class="identifier">max_digits10</span> <span class="special">=</span> <span class="number">2</span> <span class="special">+</span> <span class="special">(</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">policies</span><span class="special">::</span><span class="identifier">digits</span><span class="special"><</span><span class="keyword">double</span><span class="special">,</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">policies</span><span class="special">::</span><span class="identifier">policy</span><span class="special"><></span> <span class="special">>()</span> <span class="special">*</span> <span class="number">30103UL</span><span class="special">)</span> <span class="special">/</span> <span class="number">100000UL</span><span class="special">;</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"BOOST_NO_NUMERIC_LIMITS_LOWEST is defined"</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span>
<span class="preprocessor">#else</span>
<span class="keyword">int</span> <span class="identifier">max_digits10</span> <span class="special">=</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">numeric_limits</span><span class="special"><</span><span class="keyword">double</span><span class="special">>::</span><span class="identifier">max_digits10</span><span class="special">;</span>
<span class="preprocessor">#endif</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"Show all potentially significant decimal digits std::numeric_limits<double>::max_digits10 = "</span>
<span class="special"><<</span> <span class="identifier">max_digits10</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span>
<span class="identifier">cout</span><span class="special">.</span><span class="identifier">precision</span><span class="special">(</span><span class="identifier">max_digits10</span><span class="special">);</span> <span class="comment">// </span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">cdf</span><span class="special">(</span><span class="identifier">g05</span><span class="special">,</span> <span class="number">0.0001</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// returns 0.5000346561579232, not exact 0.5.</span>
</pre>
<p>
</p>
<p>
To get the R discrete behaviour, you simply need to round with, for example,
the <code class="computeroutput"><span class="identifier">floor</span></code> function.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">cdf</span><span class="special">(</span><span class="identifier">g05</span><span class="special">,</span> <span class="identifier">floor</span><span class="special">(</span><span class="number">0.0001</span><span class="special">))</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// returns exactly 0.5</span>
</pre>
<p>
</p>
<pre class="programlisting"><code class="computeroutput"><span class="special">></span> <span class="identifier">formatC</span><span class="special">(</span><span class="identifier">pgeom</span><span class="special">(</span><span class="number">0.9999999</span><span class="special">,</span><span class="number">0.5</span><span class="special">,</span> <span class="identifier">FALSE</span><span class="special">),</span> <span class="identifier">digits</span><span class="special">=</span><span class="number">17</span><span class="special">)</span> <span class="special">[</span><span class="number">1</span><span class="special">]</span> <span class="string">" 0.25"</span></code>
<code class="computeroutput"><span class="special">></span> <span class="identifier">formatC</span><span class="special">(</span><span class="identifier">pgeom</span><span class="special">(</span><span class="number">1.999999</span><span class="special">,</span><span class="number">0.5</span><span class="special">,</span> <span class="identifier">FALSE</span><span class="special">),</span> <span class="identifier">digits</span><span class="special">=</span><span class="number">17</span><span class="special">)[</span><span class="number">1</span><span class="special">]</span> <span class="string">" 0.25"</span> <span class="identifier">k</span> <span class="special">=</span> <span class="number">1</span></code>
<code class="computeroutput"><span class="special">></span> <span class="identifier">formatC</span><span class="special">(</span><span class="identifier">pgeom</span><span class="special">(</span><span class="number">1.9999999</span><span class="special">,</span><span class="number">0.5</span><span class="special">,</span> <span class="identifier">FALSE</span><span class="special">),</span> <span class="identifier">digits</span><span class="special">=</span><span class="number">17</span><span class="special">)[</span><span class="number">1</span><span class="special">]</span> <span class="string">"0.12500000000000003"</span> <span class="identifier">k</span> <span class="special">=</span> <span class="number">2</span></code>
</pre>
<p>
shows that R makes an arbitrary round-up decision at about 1e7 from the
next integer above. This may be convenient in practice, and could be
replicated in C++ if desired.
</p>
<h6>
<a name="math_toolkit.dist.stat_tut.weg.geometric_eg.h1"></a>
<span><a name="math_toolkit.dist.stat_tut.weg.geometric_eg.surveying_customers_to_find_one_with_a_faulty_product"></a></span><a class="link" href="geometric_eg.html#math_toolkit.dist.stat_tut.weg.geometric_eg.surveying_customers_to_find_one_with_a_faulty_product">Surveying
customers to find one with a faulty product</a>
</h6>
<p>
A company knows from warranty claims that 2% of their products will be
faulty, so the 'success_fraction' of finding a fault is 0.02. It wants
to interview a purchaser of faulty products to assess their 'user experience'.
</p>
<p>
To estimate how many customers they will probably need to contact in
order to find one who has suffered from the fault, we first construct
a geometric distribution with probability 0.02, and then chose a confidence,
say 80%, 95%, or 99% to finding a customer with a fault. Finally, we
probably want to round up the result to the integer above using the
<code class="computeroutput"><span class="identifier">ceil</span></code> function. (We could
also use a policy, but that is hardly worthwhile for this simple application.)
</p>
<p>
(This also assumes that each customer only buys one product: if customers
bought more than one item, the probability of finding a customer with
a fault obviously improves.)
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span><span class="special">.</span><span class="identifier">precision</span><span class="special">(</span><span class="number">5</span><span class="special">);</span>
<span class="identifier">geometric</span> <span class="identifier">g</span><span class="special">(</span><span class="number">0.02</span><span class="special">);</span> <span class="comment">// On average, 2 in 100 products are faulty.</span>
<span class="keyword">double</span> <span class="identifier">c</span> <span class="special">=</span> <span class="number">0.95</span><span class="special">;</span> <span class="comment">// 95% confidence.</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">" quantile(g, "</span> <span class="special"><<</span> <span class="identifier">c</span> <span class="special"><<</span> <span class="string">") = "</span> <span class="special"><<</span> <span class="identifier">quantile</span><span class="special">(</span><span class="identifier">g</span><span class="special">,</span> <span class="identifier">c</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"To be "</span> <span class="special"><<</span> <span class="identifier">c</span> <span class="special">*</span> <span class="number">100</span>
<span class="special"><<</span> <span class="string">"% confident of finding we customer with a fault, need to survey "</span>
<span class="special"><<</span> <span class="identifier">ceil</span><span class="special">(</span><span class="identifier">quantile</span><span class="special">(</span><span class="identifier">g</span><span class="special">,</span> <span class="identifier">c</span><span class="special">))</span> <span class="special"><<</span> <span class="string">" customers."</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 148</span>
<span class="identifier">c</span> <span class="special">=</span> <span class="number">0.99</span><span class="special">;</span> <span class="comment">// Very confident.</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"To be "</span> <span class="special"><<</span> <span class="identifier">c</span> <span class="special">*</span> <span class="number">100</span>
<span class="special"><<</span> <span class="string">"% confident of finding we customer with a fault, need to survey "</span>
<span class="special"><<</span> <span class="identifier">ceil</span><span class="special">(</span><span class="identifier">quantile</span><span class="special">(</span><span class="identifier">g</span><span class="special">,</span> <span class="identifier">c</span><span class="special">))</span> <span class="special"><<</span> <span class="string">" customers."</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 227</span>
<span class="identifier">c</span> <span class="special">=</span> <span class="number">0.80</span><span class="special">;</span> <span class="comment">// Only reasonably confident.</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"To be "</span> <span class="special"><<</span> <span class="identifier">c</span> <span class="special">*</span> <span class="number">100</span>
<span class="special"><<</span> <span class="string">"% confident of finding we customer with a fault, need to survey "</span>
<span class="special"><<</span> <span class="identifier">ceil</span><span class="special">(</span><span class="identifier">quantile</span><span class="special">(</span><span class="identifier">g</span><span class="special">,</span> <span class="identifier">c</span><span class="special">))</span> <span class="special"><<</span> <span class="string">" customers."</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 79</span>
</pre>
<p>
</p>
<h6>
<a name="math_toolkit.dist.stat_tut.weg.geometric_eg.h2"></a>
<span><a name="math_toolkit.dist.stat_tut.weg.geometric_eg.basket_ball_shooters"></a></span><a class="link" href="geometric_eg.html#math_toolkit.dist.stat_tut.weg.geometric_eg.basket_ball_shooters">Basket
Ball Shooters</a>
</h6>
<p>
According to Wikipedia, average pro basket ball players get <a href="http://en.wikipedia.org/wiki/Free_throw" target="_top">free
throws</a> in the baskets 70 to 80 % of the time, but some get as
high as 95%, and others as low as 50%. Suppose we want to compare the
probabilities of failing to get a score only on the first or on the fifth
shot? To start we will consider the average shooter, say 75%. So we construct
a geometric distribution with success_fraction parameter 75/100 = 0.75.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span><span class="special">.</span><span class="identifier">precision</span><span class="special">(</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">geometric</span> <span class="identifier">gav</span><span class="special">(</span><span class="number">0.75</span><span class="special">);</span> <span class="comment">// Shooter averages 7.5 out of 10 in the basket.</span>
</pre>
<p>
</p>
<p>
What is probability of getting 1st try in the basket, that is with no
failures?
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"Probability of score on 1st try = "</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">gav</span><span class="special">,</span> <span class="number">0</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.75</span>
</pre>
<p>
</p>
<p>
This is, of course, the success_fraction probability 75%. What is the
probability that the shooter only scores on the fifth shot? So there
are 5-1 = 4 failures before the first success.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"Probability of score on 5th try = "</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">gav</span><span class="special">,</span> <span class="number">4</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.0029</span>
</pre>
<p>
</p>
<p>
Now compare this with the poor and the best players success fraction.
We need to constructing new distributions with the different success
fractions, and then get the corresponding probability density functions
values:
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">geometric</span> <span class="identifier">gbest</span><span class="special">(</span><span class="number">0.95</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"Probability of score on 5th try = "</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">gbest</span><span class="special">,</span> <span class="number">4</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 5.9e-6</span>
<span class="identifier">geometric</span> <span class="identifier">gmediocre</span><span class="special">(</span><span class="number">0.50</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"Probability of score on 5th try = "</span> <span class="special"><<</span> <span class="identifier">pdf</span><span class="special">(</span><span class="identifier">gmediocre</span><span class="special">,</span> <span class="number">4</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.031</span>
</pre>
<p>
</p>
<p>
So we can see the very much smaller chance (0.000006) of 4 failures by
the best shooters, compared to the 0.03 of the mediocre.
</p>
<h6>
<a name="math_toolkit.dist.stat_tut.weg.geometric_eg.h3"></a>
<span><a name="math_toolkit.dist.stat_tut.weg.geometric_eg.estimating_failures"></a></span><a class="link" href="geometric_eg.html#math_toolkit.dist.stat_tut.weg.geometric_eg.estimating_failures">Estimating
failures</a>
</h6>
<p>
Of course one man's failure is an other man's success. So a fault can
be defined as a 'success'.
</p>
<p>
If a fault occurs once after 100 flights, then one might naively say
that the risk of fault is obviously 1 in 100 = 1/100, a probability of
0.01.
</p>
<p>
This is the best estimate we can make, but while it is the truth, it
is not the whole truth, for it hides the big uncertainty when estimating
from a single event. "One swallow doesn't make a summer." To
show the magnitude of the uncertainty, the geometric (or the negative
binomial) distribution can be used.
</p>
<p>
If we chose the popular 95% confidence in the limits, corresponding to
an alpha of 0.05, because we are calculating a two-sided interval, we
must divide alpha by two.
</p>
<p>
</p>
<pre class="programlisting"><span class="keyword">double</span> <span class="identifier">alpha</span> <span class="special">=</span> <span class="number">0.05</span><span class="special">;</span>
<span class="keyword">double</span> <span class="identifier">k</span> <span class="special">=</span> <span class="number">100</span><span class="special">;</span> <span class="comment">// So frequency of occurence is 1/100.</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"Probability is failure is "</span> <span class="special"><<</span> <span class="number">1</span><span class="special">/</span><span class="identifier">k</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span>
<span class="keyword">double</span> <span class="identifier">t</span> <span class="special">=</span> <span class="identifier">geometric</span><span class="special">::</span><span class="identifier">find_lower_bound_on_p</span><span class="special">(</span><span class="identifier">k</span><span class="special">,</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"geometric::find_lower_bound_on_p("</span> <span class="special"><<</span> <span class="keyword">int</span><span class="special">(</span><span class="identifier">k</span><span class="special">)</span> <span class="special"><<</span> <span class="string">", "</span> <span class="special"><<</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span> <span class="special"><<</span> <span class="string">") = "</span>
<span class="special"><<</span> <span class="identifier">t</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.00025</span>
<span class="identifier">t</span> <span class="special">=</span> <span class="identifier">geometric</span><span class="special">::</span><span class="identifier">find_upper_bound_on_p</span><span class="special">(</span><span class="identifier">k</span><span class="special">,</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"geometric::find_upper_bound_on_p("</span> <span class="special"><<</span> <span class="keyword">int</span><span class="special">(</span><span class="identifier">k</span><span class="special">)</span> <span class="special"><<</span> <span class="string">", "</span> <span class="special"><<</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span> <span class="special"><<</span> <span class="string">") = "</span>
<span class="special"><<</span> <span class="identifier">t</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.037</span>
</pre>
<p>
</p>
<p>
So while we estimate the probability is 0.01, it might lie between 0.0003
and 0.04. Even if we relax our confidence to alpha = 90%, the bounds
only contract to 0.0005 and 0.03. And if we require a high confidence,
they widen to 0.00005 to 0.05.
</p>
<p>
</p>
<pre class="programlisting"><span class="identifier">alpha</span> <span class="special">=</span> <span class="number">0.1</span><span class="special">;</span> <span class="comment">// 90% confidence.</span>
<span class="identifier">t</span> <span class="special">=</span> <span class="identifier">geometric</span><span class="special">::</span><span class="identifier">find_lower_bound_on_p</span><span class="special">(</span><span class="identifier">k</span><span class="special">,</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"geometric::find_lower_bound_on_p("</span> <span class="special"><<</span> <span class="keyword">int</span><span class="special">(</span><span class="identifier">k</span><span class="special">)</span> <span class="special"><<</span> <span class="string">", "</span> <span class="special"><<</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span> <span class="special"><<</span> <span class="string">") = "</span>
<span class="special"><<</span> <span class="identifier">t</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.0005</span>
<span class="identifier">t</span> <span class="special">=</span> <span class="identifier">geometric</span><span class="special">::</span><span class="identifier">find_upper_bound_on_p</span><span class="special">(</span><span class="identifier">k</span><span class="special">,</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"geometric::find_upper_bound_on_p("</span> <span class="special"><<</span> <span class="keyword">int</span><span class="special">(</span><span class="identifier">k</span><span class="special">)</span> <span class="special"><<</span> <span class="string">", "</span> <span class="special"><<</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span> <span class="special"><<</span> <span class="string">") = "</span>
<span class="special"><<</span> <span class="identifier">t</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.03</span>
<span class="identifier">alpha</span> <span class="special">=</span> <span class="number">0.01</span><span class="special">;</span> <span class="comment">// 99% confidence.</span>
<span class="identifier">t</span> <span class="special">=</span> <span class="identifier">geometric</span><span class="special">::</span><span class="identifier">find_lower_bound_on_p</span><span class="special">(</span><span class="identifier">k</span><span class="special">,</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"geometric::find_lower_bound_on_p("</span> <span class="special"><<</span> <span class="keyword">int</span><span class="special">(</span><span class="identifier">k</span><span class="special">)</span> <span class="special"><<</span> <span class="string">", "</span> <span class="special"><<</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span> <span class="special"><<</span> <span class="string">") = "</span>
<span class="special"><<</span> <span class="identifier">t</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 5e-005</span>
<span class="identifier">t</span> <span class="special">=</span> <span class="identifier">geometric</span><span class="special">::</span><span class="identifier">find_upper_bound_on_p</span><span class="special">(</span><span class="identifier">k</span><span class="special">,</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span><span class="special">);</span>
<span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"geometric::find_upper_bound_on_p("</span> <span class="special"><<</span> <span class="keyword">int</span><span class="special">(</span><span class="identifier">k</span><span class="special">)</span> <span class="special"><<</span> <span class="string">", "</span> <span class="special"><<</span> <span class="identifier">alpha</span><span class="special">/</span><span class="number">2</span> <span class="special"><<</span> <span class="string">") = "</span>
<span class="special"><<</span> <span class="identifier">t</span> <span class="special"><<</span> <span class="identifier">endl</span><span class="special">;</span> <span class="comment">// 0.052</span>
</pre>
<p>
</p>
<p>
In real life, there will usually be more than one event (fault or success),
when the negative binomial, which has the neccessary extra parameter,
will be needed.
</p>
<p>
As noted above, using a catch block is always a good idea, even if you
hope not to use it!
</p>
<p>
</p>
<pre class="programlisting"><span class="special">}</span>
<span class="keyword">catch</span><span class="special">(</span><span class="keyword">const</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">exception</span><span class="special">&</span> <span class="identifier">e</span><span class="special">)</span>
<span class="special">{</span> <span class="comment">// Since we have set an overflow policy of ignore_error,</span>
<span class="comment">// an overflow exception should never be thrown.</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">cout</span> <span class="special"><<</span> <span class="string">"\nMessage from thrown exception was:\n "</span> <span class="special"><<</span> <span class="identifier">e</span><span class="special">.</span><span class="identifier">what</span><span class="special">()</span> <span class="special"><<</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">endl</span><span class="special">;</span>
</pre>
<p>
</p>
<p>
For example, without a ignore domain error policy, if we asked for
</p>
<pre class="programlisting"><span class="identifier">pdf</span><span class="special">(</span><span class="identifier">g</span><span class="special">,</span> <span class="special">-</span><span class="number">1</span><span class="special">)</span></pre>
<p>
for example, we would get an unhelpful abort, but with a catch:
</p>
<pre class="programlisting">Message from thrown exception was:
Error in function boost::math::pdf(const exponential_distribution<double>&, double):
Number of failures argument is -1, but must be >= 0 !
</pre>
<p>
See full source C++ of this example at <a href="../../../../../../../example/geometric_examples.cpp" target="_top">geometric_examples.cpp</a>
</p>
<p>
<a class="link" href="neg_binom_eg/neg_binom_conf.html" title="Calculating Confidence Limits on the Frequency of Occurrence for the Negative Binomial Distribution">See
negative_binomial confidence interval example.</a>
</p>
</div>
<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
<td align="left"></td>
<td align="right"><div class="copyright-footer">Copyright © 2006-2010 John Maddock, Paul A. Bristow, Hubert Holin, Xiaogang Zhang, Bruno
Lalande, Johan Råde, Gautam Sewani, Thijs van den Berg and Benjamin Sobotta<p>
Distributed under the Boost Software License, Version 1.0. (See accompanying
file LICENSE_1_0.txt or copy at <a href="http://www.boost.org/LICENSE_1_0.txt" target="_top">http://www.boost.org/LICENSE_1_0.txt</a>)
</p>
</div></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="binom_eg/binom_size_eg.html"><img src="../../../../../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../weg.html"><img src="../../../../../../../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../../../../index.html"><img src="../../../../../../../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="neg_binom_eg.html"><img src="../../../../../../../../../doc/src/images/next.png" alt="Next"></a>
</div>
</body>
</html>
|
