-
Notifications
You must be signed in to change notification settings - Fork 0
/
chapter12_theory.lyx
299 lines (255 loc) · 7.38 KB
/
chapter12_theory.lyx
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
#LyX 2.3 created this file. For more info see http://www.lyx.org/
\lyxformat 544
\begin_document
\begin_header
\save_transient_properties true
\origin unavailable
\textclass amsbook
\use_default_options true
\begin_modules
theorems-ams
eqs-within-sections
figs-within-sections
\end_modules
\maintain_unincluded_children false
\language english
\language_package default
\inputencoding auto
\fontencoding global
\font_roman "default" "default"
\font_sans "default" "default"
\font_typewriter "default" "default"
\font_math "auto" "auto"
\font_default_family default
\use_non_tex_fonts false
\font_sc false
\font_osf false
\font_sf_scale 100 100
\font_tt_scale 100 100
\use_microtype false
\use_dash_ligatures true
\graphics default
\default_output_format default
\output_sync 0
\bibtex_command default
\index_command default
\paperfontsize default
\spacing single
\use_hyperref false
\papersize default
\use_geometry false
\use_package amsmath 1
\use_package amssymb 1
\use_package cancel 1
\use_package esint 1
\use_package mathdots 1
\use_package mathtools 1
\use_package mhchem 1
\use_package stackrel 1
\use_package stmaryrd 1
\use_package undertilde 1
\cite_engine basic
\cite_engine_type default
\biblio_style plain
\use_bibtopic false
\use_indices false
\paperorientation portrait
\suppress_date false
\justification true
\use_refstyle 1
\use_minted 0
\index Index
\shortcut idx
\color #008000
\end_index
\secnumdepth 3
\tocdepth 3
\paragraph_separation indent
\paragraph_indentation default
\is_math_indent 0
\math_numbering_side default
\quotes_style english
\dynamic_quotes 0
\papercolumns 1
\papersides 1
\paperpagestyle default
\tracking_changes false
\output_changes false
\html_math_output 0
\html_css_as_file 0
\html_be_strict false
\end_header
\begin_body
\begin_layout Chapter*
12.1 What is Simulation
\end_layout
\begin_layout Standard
Simulation is a way to use high-speed computer to substitute for analytical
calculation.
The law of large number tells us that if we observe a large sample of i.i.d.
random variables with finite mean, then the average of these random variables
should be close to their mean.
If we can get a computer to produce such a large i.i.d.
sample, then we can average the random variables instead of trying (and
possibly failing) to calculate their mean analytically.
For a specific problem, one needs to figure out what types of random variables
one needs, how to make a computer produce them, and how many one needs
in order to have any confidence in the numerical result.
Each of these issues will be addressed to some extent in this chapter.
\end_layout
\begin_layout Section*
Proof of Concept
\end_layout
\begin_layout Standard
We begin with some simple examples of simulation to answer questions that
we can already answer analytically just to show that simulation does what
it advertises.
Also, these simple examples will raise some of the issues to which one
must attend then trying to answer more difficult questions using simulation.
\end_layout
\begin_layout Section*
Examples in which Simulation Might Help
\end_layout
\begin_layout Standard
Next, we present some examples where the basic questions are relatively
simple to describe, by analytic solution would be tedious at best.
\end_layout
\begin_layout Section*
Summary
\end_layout
\begin_layout Standard
Suppose that we want to know the mean of some function
\begin_inset Formula $g$
\end_inset
of a random variable or random vector
\begin_inset Formula $W$
\end_inset
.
For instance, in Example
\begin_inset Formula $12.1.3$
\end_inset
we can let
\begin_inset Formula $W=(X,Y)$
\end_inset
and
\begin_inset Formula $g(W)=|X-Y|$
\end_inset
.
If a computer can supply a large number of i.i.d.
random variables (or random vectors) with the distribution of
\begin_inset Formula $W$
\end_inset
, one can use the average of the simulated values of
\begin_inset Formula $g(W)$
\end_inset
to approximate the mean of
\begin_inset Formula $g(X)$
\end_inset
.
One muse be careful to take the variability in
\begin_inset Formula $g(W)$
\end_inset
into account when deciding how much confidence to place in the approximation.
\end_layout
\begin_layout Chapter*
12.2 Why is Simulation Useful ?
\end_layout
\begin_layout Standard
Statistical simulations are used to estimate features of distributions such
as means of functions, quantiles, and other features that we cannot compute
in closed form.
When using a simulation estimator, it is good to compute a measure of how
precise the estimator is, in addition to the estimate itself.
\end_layout
\begin_layout Section*
Examples of Simulation
\end_layout
\begin_layout Section*
Assessing Uncertainty about Simulation Results
\end_layout
\begin_layout Standard
We should always attempt to assess the uncertainty in a simulation.
This uncertainty is most easily assessed via the simulation variance of
the simulated quantity.
The smaller the simulation variance is, the more certain we can be that
our estimator
\begin_inset Formula $Z$
\end_inset
is close to what we are trying to estimate.
But we need to estimate or bound the simulation before we can assess the
amount of uncertainty.
How we estimate the simulation variance of a result
\begin_inset Formula $Z$
\end_inset
depends on whether
\begin_inset Formula $Z$
\end_inset
is an average of simulated values.
The square root of our estimate of the simulation variance will be called
the simulation standard error, and it is estimate of the simulation standard
deviation of
\begin_inset Formula $Z$
\end_inset
.
The simulation standard error is a popular way to summarize uncertainty
about a simulation for two reasons.
Firsts, it has the same units of measurement as the quantity that was estimated
(unlike the simulation variance).
Second, the simulation standard error is useful for saying how likely it
is that the simulation estimator is close to the parameter being estimated.
We shall explain this second point in more detail after we show how to
calculate the simulation standard error in several common cases.
\end_layout
\begin_layout Section*
Summary
\end_layout
\begin_layout Standard
If we wish to compute the expected value
\begin_inset Formula $\theta$
\end_inset
of some random variable
\begin_inset Formula $Y$
\end_inset
, but cannot perform the necessary calculation in closed form, we can use
simulation.
In general, we would simulate a large random sample
\begin_inset Formula $Y^{(1)},...,Y^{(v)}$
\end_inset
from the same distribution as
\begin_inset Formula $Y$
\end_inset
, and then compute the sample mean
\begin_inset Formula $Z$
\end_inset
as our estimator.
We can also estimate a quantile
\begin_inset Formula $\theta_{p}$
\end_inset
of a distribution in a similar fashion.
If
\begin_inset Formula $Y^{(1)},...,Y^{(v)}$
\end_inset
is a large sample from the distribution, we can compute the sample
\begin_inset Formula $p$
\end_inset
quantile
\begin_inset Formula $Z$
\end_inset
.
It is always a good idea to compute some measure of how good a simulation
estimator is.
One common measure is the simulation standard error of
\begin_inset Formula $Z$
\end_inset
, an estimate of the standard deviation of the simulation distribution of
\begin_inset Formula $Z$
\end_inset
.
Alternatively, one could perform enough simulations to make sure that the
probability is high that the
\begin_inset Formula $Z$
\end_inset
is close to the parameter being estimated.
\end_layout
\end_body
\end_document