A
abbreviations 25-
absolute values 41, 59, 284, 422
-
accuracy
-
- numerical 116, 230-231, 423-425
-
- symbolic 421-422
-
admissible estimator 302
-
Ali-Mikhail-Haq 212, 249
-
ancillary statistic 337
-
animations
-
- approximation error 286
-
- bivariate Exponential pdf (Gumbel Model II) 11
-
- bivariate Gamma pdf (McKay) 248
-
- bivariate Normal pdf 217
-
- bivariate Normal quantiles 219
-
- bivariate Normal-Uniform pdf 214
-
- bivariate Uniform pdf 213
-
- conditional mean and variance 215
-
- contours of bivariate Normal component-mix 249
-
- contours of the trivariate Normal pdf 227
-
- limit distribution of Binomial is Poisson 281
-
- Lorenz curve for a Pareto distribution 44
-
- non-parametric kernel density estimate 183
-
- Pearson system 150
-
- pmf of sum of two dice (fair vs shaved) 87
-
- Robin Hood 223
-
arbitrary-precision numbers 423-424
-
Arc-Sine distribution 6
-
ARCH model 384-392
-
assumptions technology 8-9
-
asymptotic distribution 282-286
-
- definition 282
-
- of MLE (invariance property) 369-371
-
- of MLE (maximum likelihood estimator) 367
-
- of MLE (multiple parameters) 371-374
-
- of MLE (with hypothesis testing) 393-394
-
- of sample mean 287
-
- of sample sum 287
-
asymptotic Fisher Information 375, 376
-
asymptotic theory 277-300
-
asymptotic unbiased estimator 366
-
asymptotic variance-covariance matrix 395-399, 404, 407, 410, 415, 418-419
-
augmented symmetric function 272-276
-
Azzalini's skew-Normal distribution 80, 225
-
bandwidth 181-
Bates's distribution 139, 289-290
-
Bernoulli distribution 89-91
-
- cumulant generating function 271
-
- distribution of sample sum 141
-
- likelihood 352
-
- Logit model 90-91
-
- method of moments estimator 184
-
- pmf 89
-
- sample mean vs sample median 309-310
-
- sufficiency in Bernoulli trials 337
-
Berry-Esseen Theorem 453
-
best unbiased estimator (BUE) 325, 335-336, 362, 364
-
Beta distribution
-
- as defining Pearson Type I(J) 185
-
- as member of Pearson family 158
-
- cumulants 64
-
- fitted to censored marks 353-354
-
- MLE 363
-
- pdf 64
-
Beta-Binomial distribution 106
-
bias 306
-
Binomial distribution 91-95
-
- as limiting distribution of Ehrenfest urn 95
-
- as sum of n Bernoulli rv's 91, 141
-
- cdf 92
-
- kurtosis 93
-
- limit distribution 280, 281
-
- mgf 141, 281
-
- Normal approximation 93, 281, 299
-
- pmf 91
-
- Poisson approximation 95, 280, 300
-
- product cumulant 270
-
biology 107, 380
-
Birnbaum-Saunders distribution
-
- cdf, pdf, quantiles 38-39
-
- pseudo-random number generation 78
-
bivariate Cauchy distribution 237
-
bivariate Exponential distribution
-
- Gumbel Model I, 204
-
- Gumbel Model II, 11-13
-
bivariate Gamma (McKay) 248
-
bivariate Logistic distribution (Gumbel) 248, 249
-
bivariate Normal distribution 216-226
-
- cdf 216, 217, 229-231
-
bivariate Normal distribution (cont.)
-
- characteristic function 221
-
- component-mixture 249
-
- conditional distribution 220
-
- contour plot 218
-
- marginal distributions 220
-
- mgf 220
-
- orthant probability 231
-
- pdf 216, 217
-
- pseudo-random number generation 232-234
-
- quantiles 218-219
-
- truncated bivariate Normal 224-226
-
- variance-covariance matrix 220
-
- visualising random data 234
-
bivariate Normal-Uniform distribution 213-215
-
bivariate Poisson 243-248
-
- mgf 246
-
- moments 246-248
-
- pgf 244
-
- pmf 244-245
-
bivariate Student's t 237-238
-
bivariate Uniform (à la Morgenstern) 212-213
-
Black-Scholes option pricing 70-71, 447
-
Brownian motion 70
-
-
C
Cauchy distribution-
- as a stable distribution 58
-
- as ratio of two Normals 134
-
- as transformation of Uniform 119
-
- characteristic function 143
-
- compared to Sinc2 pdf 35-36
-
- distribution of sample mean 143
-
- mean 36
-
- pdf 35, 143
-
- product of two Cauchy rv's 148
-
cdf (cumulative distribution function)
-
- definitions
-
- - continuous multivariate 191
-
- - continuous univariate 31
-
- - discrete multivariate 194
-
- - discrete univariate 81
-
- limit distribution 279
-
- numerical cdf 39
-
- of Arc-Sine 7
-
- of Binomial 92
-
- of Birnbaum-Saunders 39
-
- of bivariate Exponential
-
- - Gumbel Model I, 204
-
- - Gumbel Model II, 12
-
- of bivariate Normal 216, 217, 229-231
-
- of bivariate Normal-Uniform 214
-
- of bivariate Uniform 213
-
- of half-Halo 75
-
- of Inverse Triangular 13
-
- of Levy 74
-
-
- of Maxwell-Boltzmann 32
-
- of Pareto distribution 38
-
- of Pascal 10
-
- of Reflected Gamma 33
-
- of stable distribution 59
-
- of trivariate Normal 229-231
-
- see also inverse cdf
-
censored data 354
-
censored distribution 68-69
-
- and option pricing 70-71
-
- and pseudo-random number generation 114
-
- censored Lognormal 71
-
- censored Normal 69
-
- censored Poisson 327
-
Central Limit Theorem 286-292, 365
-
- Generalised Central Limit Theorem 56
-
- Lindeberg-Feller 453
-
- Lindeberg-Lévy 287, 366, 368, 373
-
central moment 45, 200
-
characteristic function 50-60
-
- definition
-
- - multivariate 203
-
- - univariate 50
-
- inversion of cf
-
- - numerical 53, 55, 60
-
- - symbolic 53-60
-
- Inversion Theorem 53
-
- of bivariate Normal 221
-
- of Cauchy 58, 143
-
- of Levy 58
-
- of Lindley 51
-
- of Linnik 54
-
- of Normal 50, 57
-
- of Pareto 51
-
- of stable distribution 56-57
-
- relation to pgf 84
-
- transformations 131
-
- Uniqueness Theorem 52
-
Chebyshev's Inequality 295-296
-
Chi-squared distribution
-
- as square of a Normal rv 129, 131, 299
-
- asymptotic distribution of sample mean 283
-
- distribution of sample sum 142
-
- mean deviation 41, 421-422
-
- method of moments estimator 283
-
- mgf 131
-
- mode 36
-
- pdf 36, 41
-
- ratio of two Chi-squared rv's 135
-
- relation to Fisher F 135
-
- van Beek's bound 284-285
-
- see also noncentral Chi-squared
-
coefficient of variation 40
-
complete sufficient statistic 343, 346
-
component-mix distribution 102-104
-
- bivariate Normal component-mixture 249
-
- estimating a Poisson two-component-mix 405-411
-
-
conditional expectation E[X½a < X £ b] 66-67
-
- odd-valued Poisson rv 97-98
-
- truncated Normal 67
-
conditional expectation E[X½Y = y] 197-199
-
- definitions: continuous 197, discrete 199
-
- deriving conditional mean and variance
-
- - continuous 198, 215
-
- - discrete 199
-
- Normal Linear Regression model 221-222
-
- Rao-Blackwell Theorem 342
-
- regression function 197, 221-222
-
conditional pdf f (X½a < X £ b) 65-67
-
conditional pdf f (X½Y = y) 197
-
- of bivariate Exponential (Gumbel Model II) 12
-
- of bivariate Normal 220
-
- of bivariate Normal-Uniform 215
-
- Normal Linear Regression model 221-222
-
conditional pmf f (X½Y = y) 199
-
conditional probability 65, 97
-
confidence interval 394-395
-
consistency 292-294, 367, 457
-
consistent estimator 294, 297
-
Continuous Mapping Theorem 366, 456
-
contour plot 188, 218, 227
-
convergence
-
- in distribution 278-282, 293
-
- in probability 292-298
-
- to a constant 294
-
copulae 211-215
-
correlation 201
-
- and independence 125, 211
-
- and positive definite matrix 228
-
- between k-statistics 268
-
- between order statistics 314
-
- definition 201
-
- trivariate example 202
-
- visualising correlation 212-213
-
- see also covariance
-
covariance 201
-
- between sample moments 266
-
- definition 201
-
- derived from central mgf 205
-
- in terms of raw moments 206
-
- of bivariate Exponential (Gumbel Model II) 12
-
- trivariate example 202
-
- see also correlation
-
Cramér-Rao lower bound 333-335
-
- for Extreme Value 336
-
- for Inverse Gaussian 334-335
-
- for Poisson 334
-
cumulant generating function
-
- definition 60, 203
-
- of Bernoulli 271
-
- of Beta 64
-
- of Poisson 96
-
cumulants 60
-
- in terms of moments 62, 206-207
-
- of Bernoulli 271
-
- of Beta 64
-
- of k-statistics 267-271
-
- of Poisson 96
-
- product cumulant 209-210, 269
-
- unbiased estimator of cumulants 256-260
-
cumulative distribution function (see cdf)
-
D
data-
- censored 354
-
- population vs sample 151
-
- raw vs grouped 151
-
-
- American NFL matches 260
-
- Australian age profile 239
-
- Bank of Melbourne share price 384
-
- censored student marks 354
-
- death notices 405
-
- grain 153
-
- income and education 396
-
- medical patients and dosage 90
-
- NB1, NB2 418
-
- nerve (biometric) 380, 418
-
- psychiatric (suicide) 412
-
- sickness 155
-
- snowfall 181
-
- student marks 151, 162, 170, 177, 354
-
- Swiss bank notes 19, 185
-
- US stock market returns 185
-
- word count 418
-
degenerate distribution 103, 238, 280
-
delta method 456
-
density estimation
-
- Gram-Charlier 175-180
-
- Johnson 164-174
-
- non-parametric kernel density 181-183
-
- Pearson 149-163
-
dice 84-87
-
differentiation with respect to powers 326
-
Discrete Uniform distribution 115
-
distributions
-
- asymptotic
-
- censored
-
- component-mix
-
- degenerate
-
- elliptical
-
- empirical
-
- limit distribution
-
- mixing
-
- parameter-mix
-
- piecewise
-
- spherical
-
- stable family
-
- stopped-sum
-
- truncated
-
- zero-inflated
-
distributions - Continuous
-
- a-Laplace (see Linnik)
-
- Arc-Sine
-
- Azzalini's skew-Normal
-
- Bates
-
- Beta
-
- Birnbaum-Saunders
-
- Cauchy
-
- Chi-squared
-
- Double Exponential (see Laplace)
-
- Exponential
-
- Extreme Value
-
- Fisher F
-
- Gamma
-
- Gaussian (see Normal)
-
- half-Halo
-
- half-Normal
-
- Hyperbolic Secant
-
- Inverse Gamma
-
- Inverse Gaussian
-
- Inverse Triangular
-
- Irwin-Hall
-
- Johnson family
-
- Laplace
-
- Levy
-
- Lindley
-
- Linnik
-
- Logistic
-
- Lognormal
-
- Maxwell-Boltzmann
-
- noncentral Chi-squared
-
- noncentral F
-
- Normal
-
- Pareto
-
- Pearson family
-
- Power Function
-
- Random Walk
-
- Rayleigh
-
- Rectangular (see Uniform)
-
- Reflected Gamma
-
- semi-Circular (see half-Halo)
-
- Sinc2
-
- stable
-
- Student's t
-
- Triangular
-
- Uniform
-
- Weibull
-
distributions - Discrete
-
- Bernoulli
-
- Beta-Binomial
-
- Binomial
-
- Discrete Uniform
-
- Geometric
-
- Holla
-
- Hypergeometric
-
- Logarithmic
-
- Negative Binomial
-
- Pascal
-
- Poisson
-
- Pólya-Aeppli
-
- Riemann Zeta
-
- Waiting-time Negative Binomial
-
- Waring
-
- Yule
-
- Zero-Inflated Poisson
-
- Zipf (see Riemann Zeta)
-
distributions - Multivariate
-
- bivariate Cauchy
-
- bivariate Exponential (Gumbel Model I and II)
-
- bivariate Gamma (McKay)
-
- bivariate Logistic (Gumbel)
-
- bivariate Normal
-
- bivariate Normal-Uniform (à la Morgenstern)
-
- bivariate Poisson
-
- bivariate Student's t
-
- bivariate Uniform (à la Morgenstern)
-
- Multinomial
-
- multivariate Cauchy
-
- multivariate Gamma (Cheriyan and Ramabhadran)
-
- multivariate Normal
-
- multivariate Student's t
-
- Trinomial
-
- trivariate Normal
-
- truncated bivariate Normal
-
domain of support 31, 81-85
-
- circular 191
-
- non-rectangular 124, 125, 190-191, 314
-
- rectangular 124, 190
-
- triangular 191, 314, 317
-
dominant estimator 302
-
Dr Faustus 421
-
E
economics and finance 43-45, 56, 70-72, 108-109, 117, 121, 384-
Ehrenfest urn 94-95
-
ellipse 218, 236
-
ellipsoid 227
-
elliptical distributions 234
-
empirical pdf 73, 77, 154, 381, 383
-
empirical pmf 16, 110, 111, 112
-
engineering 122
-
entropy 15
-
Epanechnikov kernel 182
-
estimator
-
- admissible 302
-
- asymptotic unbiased 366
-
- BUE (best unbiased) 325, 335-336, 362, 364
-
- consistent 294, 297
-
- density (see density estimation)
-
- dominant 302
-
- estimator vs estimate 357
-
- Fisher estimator 395-396, 397, 404
-
- h-statistic 253-256
-
- Hessian estimator 395-396, 398, 404
-
- inadmissible 302, 321-322
-
- k-statistic 256-261
-
- maximum likelihood estimator (see MLE)
-
- method of moments 183-184, 283
-
- minimax 305
-
- minimum variance unbiased 341-346, 364
-
- non-parametric kernel density 181-183
-
- ordinary least squares 385
-
- Outer-product 395-396, 398
-
- sample central moment 360
-
- sample maximum 320-321
-
- sample mean (see sample mean)
-
- sample median 309-310, 318-320
-
- sample range 320-321
-
- sample sum 277, 287
-
- unbiased estimator of parameters 325-347
-
- unbiased estimator of population moments 251-261
-
expectation operator
-
- basic properties 32
-
- definitions
-
- - continuous 32
-
- - discrete 83
-
- - multivariate 200
-
- when applied to sample moments 263
-
Exponential distribution
-
- bivariate 11-13, 204
-
- difference of two Exponentials 139-140
-
- distribution of sample sum 141-142
-
- likelihood 351
-
- MLE (numerical) 381
-
- MLE (symbolic) 358
-
- order statistics 313-314
-
- pdf 141, 313, 344, 358
-
- relation to Extreme Value 121
-
- relation to Pareto 121
-
- relation to Rayleigh 122
-
- relation to Uniform 121
-
- sufficient statistic 344
-
- sum of two Exponentials 136
-
Exponential regression 375-376, 396
-
Extreme Value distribution
-
- Cramér-Rao lower bound 336
-
- pdf 336, 377
-
- relation to Exponential 121
-
F
factorial moment 60, 206-207, 247-
factorial moment generating function 60, 203, 247
-
factorisation criterion 339-341
-
families of distributions
-
- Gram-Charlier 175-180
-
- Johnson 164-174
-
- Pearson 149-163
-
- stable family 56-61
-
fat tails 56, 108-109
-
- see also kurtosis
-
first-order condition 21, 36, 357-361, 363
-
Fisher estimator 395-396, 397, 404
-
Fisher F distribution 135
-
Fisher Information 326-332
-
- and MLE (regularity conditions) 367-368, 372-373
-
- asymptotic Fisher Information 375, 376
-
- first derivative form vs second derivative 329
-
- for censored Poisson 327-328
-
- for Gamma 331-332
-
- for Inverse Gaussian 18
-
- for Lindley 326
-
- for Normal 330-331
-
- for Riemann Zeta 329
-
- for Uniform 330
-
Frank 212
-
frequency polygon 73, 77, 151, 154, 380
-
- see also plotting techniques
-
Function Form 82
-
functions of random variables 117-148
-
fundamental expectation result 274
-
G
games-
- archery (Robin Hood) 222-224
-
- cards, poker 101
-
- craps 87-89, 115
-
- dice (fair and unfair) 84-87
-
Gamma distribution
-
- as member of Pearson family 157, 185
-
- as sum of n Exponential rv's 141-142
-
- bivariate Gamma (McKay) 248
-
- Fisher Information 331-332
-
- hypothesis testing 392-394
-
- method of moments estimator 184
-
- mgf 142, 456
-
- MLE (numerical) 382-383
-
- multivariate (Cheriyan & Ramabhadran) 208
-
- pdf 73, 142
-
- pseudo-random number generation 73
-
- relation to Inverse Gamma 147
-
Gamma regression model 419
-
gas molecules 32
-
Gaussian kernel 19, 182
-
generating functions 46-56, 203-205
-
Geometric distribution
-
- definition 98
-
- distribution of difference of two rv's 148
-
- pmf 98
-
Gini coefficient 40, 43-45
-
gradient 357-361
-
Gram-Charlier expansions 175-180
-
graphical techniques (see plotting techniques)
-
Greek alphabet 28
-
H
h-statistic 253-256-
half-Halo distribution 75, 80
-
half-Normal distribution 225
-
Helmert transformation 145
-
HELP 5
-
Hermite polynomial 175, 179, 449
-
Hessian estimator 395-396, 398, 404
-
Hessian matrix 358, 360
-
histogram 18, 155 (see also plotting techniques)
-
Holla's distribution 105, 112
-
Hyperbolic Secant distribution 80
-
Hypergeometric distribution 100-101
-
I
inadmissible estimator 302, 321-322-
income distribution 43-44, 121
-
independence
-
- correlation and dependence 125, 211
-
- mutually stochastically independent 210
-
independent product space 124, 190
-
Invariance Property 360, 369-371, 401, 410, 417
-
inverse cdf
-
- numerical inversion 38-39, 75-77, 109
-
- symbolic inversion 37-38, 74-75
-
- of Birnbaum-Saunders 38-39
-
- of half-Halo 75
-
- of Levy 74
-
- of Pareto 38, 43
-
Inverse Gamma distribution
-
- as member of Pearson family 185
-
- pdf 365
-
- relation to Gamma 147, 365
-
- relation to Levy 58
-
Inverse Gaussian distribution
-
- Cramér-Rao lower bound 334-335
-
- Fisher Information 18
-
- pdf 18, 334
-
- relation to Random Walk distribution 147
-
Inverse Triangular distribution 13-14
-
Inversion Theorem 53
-
Irwin-Hall distribution 55, 139
-
isobaric 272
-
J
Jacobian of the transformation 118, 123, 130, 223-
Johnson family 164-174
-
- as transformation of a Logistic rv 185
-
- as transformation of a Normal rv 164
-
- Types and chart 164
-
- - SL (Lognormal) 165-167
-
- - SU (Unbounded) 168-172
-
- - SB (Bounded) 173-174
-
K
k-statistic 20, 256-261-
kernel density (see non-parametric kernel density)
-
Khinchine's Theorem 298
-
Khinchine's Weak Law of Large Numbers 278, 296-298, 366
-
Kronecker product 437
-
kurtosis
-
- building your own function 446
-
- definition 40-41
-
- of Binomial 93
-
- of Poisson 446
-
- of Weibull 42
-
- Pearson family 149-150
-
L
Laplace distribution-
- as Linnik 54
-
- as Reflected Gamma 33
-
- order statistics of 23, 315-317
-
- relation to Exponential 139-140
-
latent variable 353, 412
-
Lehmann-Scheffé Theorem 346
-
Levy distribution
-
- as a stable distribution 58
-
- as an Inverse Gamma 58
-
- cdf, pdf, pseudo-random number 74
-
likelihood
-
- function 21, 350-357
-
- observed 22, 351-357
-
- see also log-likelihood
-
limit distribution
-
- definition 279
-
- of Binomial 280, 281
-
- of sample mean (Normal) 279
-
limits in Mathematica 278
-
Lindley distribution
-
- characteristic function 51
-
- Fisher Information 326-327
-
- pdf 51, 327
-
linear regression function 221
-
linex (linear-exponential) loss 322
-
linguistics 107
-
Linnik distribution 54
-
List Form 82, 111
-
log-likelihood
-
- concentrated 361, 382-383, 418
-
- function 21, 357-376, 381
-
- observed log-likelihood
-
- - ARCH model (stock prices) 387
-
- - Exponential model (nerve data) 381
-
- - Exponential regression (income) 396
-
- - Gamma model (nerve data) 382-383
-
- - Logit model (dosage data) 90
-
- - Ordered Probit model (psychiatric data) 414-415
-
- - Poisson two-component-mix model 405-406
-
- see also likelihood
-
Logarithmic distribution 115
-
Logistic distribution
-
- as base for a Johnson-style family 185
-
- bivariate 248, 249
-
- pdf 23, 318
-
- order statistics of 23
-
- relation to Uniform 147
-
- sample mean vs sample median 318-320
-
Logit model 90-91
-
Lognormal distribution
-
- and stock prices 71
-
- as member of Johnson family 165-167
-
- as transformation of Normal 120, 165
-
- censored below 71
-
- moments of sample sum 276
-
- pdf 71, 120
-
Lorenz curve 43-44
-
loss function 301-305
-
- asymmetric 303-304
-
- asymmetric quadratic 322, 323
-
- linex (linear-exponential) 322
-
- quadratic 306
-
M
machine-precision numbers 423-425-
marginal distribution 195-196
-
- and copulae 211
-
- joint pdf as product of marginals 210, 211, 351, 355
-
- more examples 12, 126, 133-137, 146, 204, 214, 220, 224-225, 237-238, 244
-
Markov chain 94, 447-448
-
Markov's inequality 295-296
-
Mathematica
-
- assumptions technology 8-9
-
- bracket types 27
-
- changes to default behaviour 443-445
-
- differentiation with respect to powers 326
-
- Greek alphabet 28
-
- how to enter 30
-
- kernel (fresh and crispy) 5, 425
-
- limits 278
-
- lists 428-429
-
- matrices 433-437, 445
-
- notation (common) 27
-
- notation entry 28-30
-
- packages 425
-
- replacements 27
-
- subscripts 429-432
-
- timings 30
-
- upper and lower case conventions 24
-
- using G in Input cells 443
-
- vectors 438-443
-
- see also plotting techniques
-
mathStatica
-
- Basic vs Gold version 4
-
- Continuous distribution palette 5
-
- Discrete distribution palette 5
-
- HELP 5
-
- installation 3
-
- loading 5
-
- registration 3
-
- working with parameters 8
-
maximum likelihood estimation (see MLE)
-
Maxwell-Boltzmann distribution 32
-
mean 35-36, 45
-
- see also sample mean
-
mean deviation 40, 41, 299, 421-422
-
mean square error (see MSE)
-
median 37
-
- of Pareto distribution 37-38
-
- see also sample median
-
medical 90-91, 155, 380, 405, 412
-
method of moments estimator 183-184
-
- for Bernoulli 184
-
- for Chi-squared 283
-
- for Gamma 184
-
mgf (moment generating function)
-
- and cumulant generating function 60
-
- and independence 210
-
- central mgf 93, 203, 205, 247
-
- definition 46, 203
-
- Inversion Theorem 53
-
- Uniqueness Theorem 52
-
- of Binomial 93, 141, 281
-
- of bivariate Exponential (Gumbel Model I) 204
-
- of bivariate Exponential (Gumbel Model II) 12
-
- of bivariate Normal 220
-
- of bivariate Poisson 246
-
- of Chi-squared 131
-
- of Gamma 142, 456
-
- of Multinomial 239, 241-242, 242-243
-
- of multivariate Gamma 208
-
- of multivariate Normal 249
-
- of noncentral Chi-squared 144
-
- of Normal 47
-
- of Pareto 49
-
- of sample mean 141
-
- of sample sum 141
-
- of sample sum of squares 141
-
- of Uniform 48
-
MGF Method 52-56, 130-132, 141-147
-
MGF Theorem 52, 141
-
- more examples 281, 364-365
-
minimax estimator 305
-
minimum variance unbiased estimation (see MVUE)
-
mixing distributions 102-109
-
- component-mix 102-104, 249, 405-411
-
- parameter-mix 105-109
-
-
MLE (maximum likelihood estimation) 357-376
-
- asymptotic properties 365-366, 371-376
-
- general properties 362
-
- invariance property 369-371
-
- more than one parameter 371-374
-
- non-iid samples 374-376
-
- numerical MLE (see Chapter 12)
-
- - ARCH model (stock prices) 387
-
- - Exponential model (nerve data) 381
-
- - Exponential regression model (income) 396
-
- - Gamma model (nerve data) 382-383
-
- - Logit model (dosage data) 90
-
- - Normal model (random data) 418
-
- - Ordered Probit model (psychiatric data) 414-415
-
- - Poisson two-component-mix model 405-406
-
- regularity conditions
-
- - basic 367-369
-
- - more than one parameter 371-372
-
- - non-iid samples 374-375
-
- small sample properties 363-365
-
- symbolic MLE (see Chapter 11)
-
- - for Exponential 358
-
- - for Normal 359-360, 418
-
- - for Pareto 360-361
-
- - for Power Function 362-363
-
- - for Rayleigh 21
-
- - for Uniform 377
-
mode 36
-
moment conversion functions
-
- univariate 62-64
-
- multivariate 206-210
-
moment generating function (see mgf)
-
moments
-
- central moment 45, 200
-
- factorial moment 60, 206-207
-
- fitting moments (see Pearson, Johnson, method of moments)
-
- negative moment 80
-
- population moments vs sample moments 251
-
- product moment 200, 266
-
- raw moment 45, 200
-
moments of moments 261-271
-
- introduction 20
-
moments of sampling distributions 251-276
-
monomial symmetric function 273
-
Monte Carlo 290
-
- see also pseudo-random number generation
-
- see also simulation
-
Morgenstern 212
-
MSE (mean square error)
-
- as risk 306-311
-
- comparing h-statistics with polyaches 264-266
-
- of sample median and sample mean (Logistic) 318-320
-
- of sample range and sample maximum (Uniform) 320-321
-
- weak law of large numbers 296-297
-
multinomial coefficient 451
-
Multinomial distribution 238-243
-
multiple local optima 400
-
multivariate Cauchy distribution 236
-
multivariate Gamma distribution (Cheriyan and Ramabhadran) 208
-
multivariate Normal distribution 216-235
-
multivariate Student's t 236
-
mutually stochastically independent 210
-
MVUE (minimum variance unbiased estimation) 341-346, 364
-
N
Negative Binomial distribution 99, 105, 418-
noncentral Chi-squared distribution
-
- as Chi-squared-Poisson mixture 105
-
- derivation 144
-
- exercises 299
-
noncentral F distribution 135
-
non-parametric kernel density 181-183
-
- with bi-weight, tri-weight kernel 182
-
- with Epanechnikov kernel 182
-
- with Gaussian kernel 19, 182
-
non-rectangular domain 124, 125, 190-191, 320-321
-
Normal distribution
-
- and Gram-Charlier expansions 175
-
- as a stable distribution 57
-
- as limit distribution of a Binomial 93, 281, 299
-
- as member of Johnson family 164-165, 167
-
- as member of Pearson family 150, 158
-
- asymptotic distribution of MLE of (m, s2) 372-374
-
- basics 8
-
- bivariate Normal 216-226
-
- censored below 69
-
- central moments 265
-
- characteristic function 50, 57
-
- characteristic function of X1X2 132
-
- conditional expectation of sample median, given sample mean 342-343
-
- distribution:
-
- - of product of two Normals 132, 133
-
- - of ratio of two Normals 134
-
- - of X2 129, 131
-
- - of sample mean 143, 294-295
-
- - of sample sum of squares 144
-
- - of sample sum of squares about the mean 145
-
- estimators for the Normal variance 307-308
-
- finance 56, 108-109
-
- Fisher Information 330-331
-
- limit distribution of sample mean 279
-
- limit Normal distribution 362, 367
-
- - examples 369, 392-395
-
- mgf 47
-
- mgf of X2 131
-
- MLE of (m, s2) 359-360, 418
-
- MVUE of (m, s2) 346
-
- Normal approximation to Binomial 93, 281, 299
-
- pseudo-random number generation
-
- - approximate 291-292
-
- - exact 72-73, 418
-
- QQ plot 291
-
- raw moments 46
-
- relation to Cauchy 134
-
- relation to Chi-squared 129, 131
-
- relation to Lognormal 120
-
- risk of a Normally distributed estimator 303-304
-
- sample mean as consistent estimator of population mean 294-295
-
- standardising a Normal rv 120
-
- sufficient statistics for (m, s2) 340-341
-
- trivariate Normal 226-228
-
- truncated above 65-66, 67
-
- working with s vs s2 326, 377, 455
-
- see also Invariance Property
-
Normal linear regression model 221-222, 385, 457
-
notation
-
- Mathematica notation
-
- - bracket types 27
-
- - Greek alphabet 28
-
- - how to enter 30
-
- - notation (common) 27
-
- - notation entry 28-30
-
- - replacements 27
-
- - subscripts 429-432
-
- - upper and lower case conventions 24
-
- - using G in Input cells 443
-
- statistics notation
-
- - abbreviations 25
-
- - sets and operators 25
-
- - statistics notation 26
-
- - upper and lower case conventions 24
-
O
one-to-one transformation 118-
optimisation
-
- differentiation with respect to powers 326
-
- first-order condition 21, 36, 357-361, 363
-
- gradient 357-361
-
- Hessian matrix 358, 360
-
- multiple local optima 400
-
- score 357-361
-
- second-order condition 22, 36-37, 357-360
-
- unconstrained vs constrained numerical optimisation 369, 379, 388-389, 401, 414
-
optimisation algorithms 399-405
-
- Armijo 408
-
- BFGS (Broyden-Fletcher-Goldfarb-Shanno) 399-400, 403, 405-411, 459
-
- DFP (Davidon-Fletcher-Powell) 403
-
- direct search 400
-
- genetic 400
-
- Golden Search 401
-
- Goldstein 408
-
- gradient method 400, 401-405
-
- line search 401
-
- Method -> Newton 390-391, 397, 403, 415, 459
-
- Method -> QuasiNewton 403, 406-407, 419, 459
-
- NR (Newton Raphson) 390-391, 397, 399-400, 403, 412-417, 458-459
-
- numerical convergence 404-405
-
- Score 403-404
-
- simulated annealing 400
-
- taboo search 400
-
option pricing 70-72
-
order statistics 311-322
-
- distribution of:
-
- - sample maximum 312, 321
-
- - sample minimum 312
-
- - sample median 318-320
-
- - sample range 320-321
-
- for Exponential 313-314
-
- for Laplace 23, 315-317
-
- for Logistic 23
-
- for Uniform 312
-
- joint order statistics 23, 314, 316, 320
-
Ordered Probit model 412-417
-
ordinary least squares 385
-
orthant probability 231
-
Outer-product estimator 395-396, 398
-
P
p-value 393-394-
parameter identification problem 414
-
parameter-mix distribution 105-109
-
Pareto distribution
-
- characteristic function 51
-
- median 37-38
-
- mgf 49
-
- MLE 360-361
-
- pdf 37, 49, 51, 360
-
- quantiles 38
-
- relation to Exponential 121
-
- relation to Power Function 147
-
- relation to Riemann Zeta 107
-
Pascal distribution 10, 99
-
pdf (probability density function)
-
- definition 31, 187
-
- see also Distributions
-
- see also pmf (for discrete rv's)
-
peakedness 40-41, 108-109
-
Pearson family 149-163
-
- animated tour 150
-
- Pearson coefficients in terms of moments 159-160
-
- Types and chart 150
-
- - Type I, 17, 156, 158, 185
-
- - Type II, 158
-
- - Type III, 154, 157, 185
-
- - Type IV, 151-153, 157
-
- - Type V, 158, 185
-
- - Type VI, 158
-
- - Type VII, 157
-
- unimodal 179
-
- using a cubic polynomial 161-163
-
penalty function 400, 407, 415
-
pgf (probability generating function)
-
- definitions 60, 84, 203
-
- deriving probabilities from pgf 85, 85-86, 86, 104, 245
-
- of bivariate Poisson 244-245
-
- of Hypergeometric 100
-
- of Negative Binomial 99
-
- of Pascal 11
-
- of Zero-Inflated Poisson 104
-
physics 32, 94-95
-
piecewise distributions
-
- Bates's distribution 289-290
-
- Inverse Triangular 13
-
- Laplace 23, 315-317
-
- order statistics of 23
-
- Reflected Gamma 33
-
plotting techniques (some examples)
-
- arrows 37, 81, 280
-
- contour plots 188, 218, 227
-
- data
-
- - bivariate / trivariate 233-235
-
- - grouped data 18, 155
-
- - raw 151
-
- see also frequency polygon
-
- - scatter plot 397
-
- - time-series 384
-
- see also empirical pdf / pmf
-
- domain of support (bivariate) 125, 138, 140
-
- empirical pdf 73, 77, 154, 381, 383
-
- empirical pmf 16, 110, 111, 112
-
- filled plot 44, 68
-
- frequency polygon 73, 77, 151, 154, 380
-
- graphics array 32, 38, 68, 109, 118, 124, 168, 174, 218
-
- histogram 18, 155
-
- Johnson system 170
-
- non-parametric kernel density 19, 182-183
-
- parametric plot 167
-
- pdf plots 6, 139, etc.
-
- - as parameters change 8, 14, 32, 145, 165, 225, 313, 315
-
- - 3D 11, 188, 198, 213, 214, 217, 316
-
- Pearson system 17, 152
-
- pmf plots 10, 83, 98, 101, 103
-
- - as parameters change 87, 92, 96
-
- - 3D 190
-
- QQ plots 291
-
- scatter plot 397
-
- superimposing plots 34, 35, 37, 42, 54, 55, 69, 91, 133, 219, 302, 306
-
- text labels 32, 37, 54, 145, 302, 306, 313
-
- wireframe 228
-
- see also animations
-
pmf (probability mass function)
-
- definitions 82, 189
-
- see also Distributions - Discrete
-
- see also pdf (for continuous rv's)
-
Poisson distribution 95-98
-
- as limit distribution of Binomial 95, 280, 300
-
- bivariate Poisson 243-248
-
- censoring 327-328
-
- Cramér-Rao lower bound 334
-
- cumulant generating function 96
-
- distribution of sample sum 137
-
- kurtosis 446
-
- odd-valued Poisson 97-98
-
- pmf 16, 95, 110, 334
-
- Poisson two-component-mix 102-103, 406
-
- pseudo-random number generation 16, 110
-
- sufficient statistic for l 340
-
- zero-inflated Poisson 104
-
poker 101
-
Pólya-Aeppli distribution 105
-
polyache 255-256
-
polykay 257-259
-
Power Function distribution
-
- as a Beta rv 185, 363
-
- as defining Pearson Type I(J) 185
-
- MLE 362-363
-
- relation to Pareto 147
-
- sufficient statistic 363-364
-
power sum 252, 272-276
-
probability
-
- conditional 65, 97
-
- multivariate 191-194
-
- orthant probability 231
-
- probability content of a region 192-193, 230-231
-
- throwing a die 84-87
-
- see also cdf
-
probability density function (see pdf)
-
probability generating function (see pgf)
-
probability mass function (see pmf)
-
probit model 412-413
-
product moment 200, 266
-
products / ratios of random variables 133-136
-
- see also:
-
- - deriving the pdf of the bivariate t 237-238
-
- - product of two Uniforms 126-127
-
Proportional-hazards model 412
-
Proportional-odds model 412
-
pseudo-random number generation
-
- methods
-
- - inverse method (numerical) 75-77, 109-115
-
- - inverse method (symbolic) 74-75
-
- - Mathematica's Statistics package 72-73
-
- - rejection method 77-79
-
- and censoring 114
-
- computational efficiency 113, 115
-
- List Form 111
-
- of Birnbaum-Saunders 78
-
- of Gamma 73
-
- of half-Halo 75-77
-
- of Holla 112
-
- of Levy 74
-
- of multivariate Normal 232-234
-
- of Normal 291-292, 418
-
- of Poisson 16, 110
-
- of Riemann Zeta 113
-
- visualising random data in 2D, 3D 233-235
-
Q
QQ plot 291-
quantiles 37
-
- of Birnbaum-Saunders 38-39
-
- of bivariate Normal 218-219
-
- of bivariate Student's t 237
-
- of Pareto 38
-
- of trivariate Normal 227-228
-
R
random number (see pseudo-random number)-
random variable
-
- continuous 31, 81, 187
-
- discrete 81-82, 189
-
- see also Distributions
-
Random Walk distribution 147
-
random walk with drift 355, 384-386
-
Rao-Blackwell Theorem 342
-
raw moment 45, 200
-
Rayleigh distribution
-
- MLE 21
-
- relation to Exponential 122
-
rectangular domain 124, 190
-
reference computer 30
-
Reflected Gamma distribution 33-34
-
registration 3
-
regression 384-392
-
regression function 197, 221-222
-
regularity conditions
-
- for Fisher Information 329-330
-
- for MLE
-
- - basic 367-369
-
- - more than one parameter 371-372
-
- - non-iid samples 374-375
-
relative mean deviation 299
-
re-parameterisation 369, 388-389, 401, 406, 410, 414
-
Riemann Zeta distribution
-
- area of application 107
-
- Fisher Information 329
-
- pmf 113, 329
-
- pseudo-random number generation 113
-
risk 301-305
-
Robin Hood 222-224
-
S
sample information 332, 338, 376-
sample maximum 311, 312, 320-321, 377
-
sample mean
-
- as consistent estimator (Khinchine) 298
-
- as consistent estimator (Normal) 294-295
-
- as MLE (for Exponential parameter) 358
-
- as MLE (for Normal parameter) 359-360
-
- asymptotic distribution of sample mean 287
-
- definition 277
-
- distribution of sample mean
-
- - for Cauchy 143
-
- - for Normal 143
-
- - for Uniform 139, 288-292
-
- Khinchine's Theorem 298
-
- limit distribution of sample mean (Normal) 279
-
- mgf of 141
-
- variance of the sample mean 264
-
- vs sample median, for Bernoulli trials 309-310
-
- vs sample median, for Logistic trials 318-320
-
sample median
-
- conditional expectation of sample median, given sample mean 342-343
-
- vs sample mean, for Bernoulli trials 309-310
-
- vs sample mean, for Logistic trials 318-320
-
sample minimum 311, 312
-
sample moment 251
-
- sample central moment 251, 360
-
- - covariance between sample central moments 266
-
- - in terms of power sums 252
-
- - variance of 264
-
- sample raw moment 251
-
- - as unbiased estimators of population raw moments 253
-
- - in terms of power sums 252
-
sample range 320-321
-
sample sum
-
- asymptotic distribution of sample sum 287
-
- definition 277
-
- distribution of sample sum
-
- - for Bernoulli 141
-
- - for Chi-squared 142
-
sample sum (cont.)
-
- distribution of sample sum (cont.)
-
- - for Exponential 141-142
-
- - for Poisson 137
-
- - for Uniform 55, 139
-
- mgf of sample sum 141
-
- moments of sample sum 261-271, 276
-
sample sum of squares
-
- distribution of (Normal) 144
-
- mgf of 141
-
sampling with or without replacement 100
-
scedastic function 197
-
score 357-361
-
second-order condition 22, 36-37, 357-360
-
security (stock) price 70-72, 108-109, 384
-
Sheather-Jones optimal bandwidth 19, 182
-
signal-to-noise ratio 299
-
Silverman optimal bandwidth 182
-
simulation 87-89, 126-127, 298-299
-
- see also Monte Carlo
-
- see also pseudo-random number
-
Sinc2 distribution 35-36
-
skewness
-
- definition 40
-
- of Weibull 42
-
- Pearson family 149-150
-
Skorohod's Theorem 456
-
small sample accuracy 289-292
-
smoothing methods 181-183
-
spherical distributions 234, 451
-
stable distributions 56-61
-
standard deviation 40, 45
-
standard error 395, 399
-
standardised random variable 40, 120, 281, 287
-
statistic 251
-
stopped-sum distribution 108
-
Student's t distribution
-
- as member of Pearson family 157
-
- as Normal-InverseGamma mixture 105
-
- bivariate Student's t 237-238
-
- derivation, pdf 134
-
sufficient statistic 337-341, 344, 362, 363-364
-
sums of random variables 136-147
-
- deriving pmf of bivariate Poisson 244-245
-
- sum of Bernoulli rv's 141
-
- sum of Chi-squared rv's 142
-
- sum of Exponentials 136, 141-142
-
- sum of Poisson rv's 137
-
- sum of Uniform rv's 54-55, 138-139
-
- see also sample sum
-
Swiss bank notes 19, 185
-
symmetric function 253, 272-276
-
systems of distributions (see families) 149-180
-
T
t distribution (see Student's t)-
t-statistic 395, 399
-
theorems
-
- Berry-Esseen 453
-
- Central Limit Theorem 286-292
-
- Continuous Mapping Theorem 366, 456
-
- Inversion Theorem 53
-
- Khinchine 298
-
- Lehmann-Scheffé 346
-
- Lindeberg-Feller 453
-
- Lindeberg-Lévy 287
-
- MGF Theorem 52, 141
-
- Rao-Blackwell Theorem 342
-
- Skorohod's Theorem 456
-
- transformation theorems
-
- - univariate 118
-
- - multivariate 123
-
- - not one-to-one 127
-
- Uniqueness Theorem 52
-
timings 30
-
transformations 117-148
-
- MGF Method 52-56, 130-132, 141-147
-
- transformation method 118-130
-
- - univariate 118
-
- - multivariate 123
-
- - manual 130
-
- - Jacobian 118, 123, 130, 223
-
- - one-to-one transformation 118
-
- - not one-to-one 127
-
- Helmert transformation 145
-
- non-rectangular domain 124, 125
-
- transformation to polar co-ordinates 222-223
-
- see also:
-
- - products / ratios of random variables
-
- - sums of random variables
-
Triangular distribution
-
- as sum of two Uniform rv's 55, 138-139
-
Trinomial distribution 239
-
trivariate Normal 226-228
-
- cdf 229-231
-
- orthant probability 231
-
- pseudo-random number generation 232-234
-
- visualising random data 235
-
truncated distribution 65-67
-
- truncated (above) standard Normal 65-66, 67
-
- truncated bivariate Normal 224-226
-
U
unbiased estimators of parameters 325-347-
- asymptotic unbiasedness 366
-
unbiased estimators of population moments 251-261
-
- introduction 20
-
- multivariate 259-261
-
- of central moments 253-254, 259-261
-
- of cumulants 256-258, 260
-
- of Normal population variance 307-308
-
- of population variance 253, 254
-
- of raw moments 253
-
Uniform distribution
-
- bivariate Uniform (à la Morgenstern) 212-213
-
- Fisher Information 330
-
- mgf 48
-
- MLE 377
-
- order statistics 312
-
- other transformations of a Uniform rv 122
-
- pdf 48, 122, 312, 320, 330
-
- product of two Uniform rv's 126-127
-
- relation to Bates 139, 289-290
-
- relation to Cauchy 119
-
- relation to Exponential 121
-
- relation to Irwin-Hall 55, 139
-
- relation to Logistic 147
-
- sample mean and Central Limit Theorem 288-292
-
- sample range vs sample maximum 320-321
-
- sum of Uniform rv's 54-55, 138-139
-
unimodal 36, 179, 182-183
-
Uniqueness Theorem 52
-
V
van Beek bound 283-285, 453-
variance
-
- definition 40, 45
-
- of sample mean 264
-
- of 2nd sample central moment 264
-
variance-covariance matrix
-
- asymptotic variance-covariance matrix 395-399, 404, 407, 410, 415, 418-419
-
- definition 201
-
- of bivariate Exponential
-
- - Gumbel Model I, 205
-
- - Gumbel Model II, 12
-
- of bivariate Normal 220
-
- of bivariate Normal-Uniform 215
-
- of bivariate Uniform 213
-
- of trivariate models 202, 211
-
- of truncated bivariate Normal 226
-
- of unbiased estimators 333-335
-
W
Waiting-time Negative Binomial distribution 99-
Waring distribution 418
-
weak law of large numbers 296-298
-
Weibull distribution 42
-
X
xenium (see book cover)-
Y
Yule distribution 107-
Z
zero-inflated distributions 103-104-
Zipf distribution (see Riemann Zeta) 107
-
|