Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Applied regression analysis / Lab / 1 / cookbook-en.pdf
Скачиваний:
34
Добавлен:
10.05.2015
Размер:
1.26 Mб
Скачать

8Probability and Moment Generating Functions

GX(t) = E tX

jtj < 1

1 (Xt)i

 

1 E Xi

 

 

i

MX(t) = GX(et) = E eXt = E "

 

 

# =

 

 

 

t

 

 

i!

i=0

i!

 

 

 

 

i=0

 

 

 

 

 

X

 

X

 

 

 

 

P [X = 0] = GX(0)P [X = 1] = G0X(0)

P [X = i] = G(Xi)(0) i!

E [X] = G0X(1 )

E Xk = MX(k)(0)

E

X!

= GX(k)

(1 )

 

 

 

(X k)!

 

 

 

V

 

 

 

 

 

 

(1 ))2

[X] = G00 (1 ) + G0

(1 )

 

(G0

 

X

X

d

X

 

GX(t) = GY (t) =)

 

 

 

 

X = Y

 

 

9Multivariate Distributions

9.1Standard Bivariate Normal

Let X; Y N (0; 1) ^ X ?? Z where Y = X + p

1 2

Z

 

Joint density

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1

 

 

 

2

+ y2

 

2 xy

 

f(x; y) =

 

 

 

 

exp

 

x

 

 

 

 

 

 

 

 

2(1

2)

2

1

 

2

Conditionals

p

 

 

 

 

 

 

 

 

 

(Y j X = x) N x; 1 2

 

and

 

(X j Y = y) N y; 1 2

Independence

X ?? Y () = 0

 

 

 

 

 

 

 

9.2Bivariate Normal

Let X N x; x2 and Y N y; y2 .

 

 

exp

 

 

 

 

 

 

f(x; y) =

1

 

 

 

 

z

 

 

 

 

 

 

 

 

 

2(1

2)

 

2 x y 1

 

2

 

"

x

 

2

 

p

2

 

 

 

 

 

y

#

 

y

 

 

x

 

 

z =

x x

 

 

+

y y

 

 

 

2

x x

 

 

y y

 

 

 

 

 

 

 

 

 

 

 

 

 

Conditional mean and variance

E [X j Y ] = E [X] + X (Y E [Y ])

Y

p

V [X j Y ] = X 1 2

9.3Multivariate Normal

Covariance matrix (Precision matrix 1)

 

 

 

=

0

V [X...

1]

... Cov [X...

1; Xk]1

 

BCov [Xk; X1]

 

V [Xk]

C

If X N ( ; ),

@

 

 

 

 

A

 

 

 

2

(x )T 1(x )

fX(x) = (2 ) n=2 j j 1=2 exp

 

 

 

 

1

 

 

 

 

Properties

Z N (0; 1) ^ X = + 1=2Z =) X N ( ; )

X N ( ; ) =) 1=2(X ) N (0; 1)

X N ( ; ) =) AX N A ; A AT

X N ( ; ) ^ kak = k =) aT X N aT ; aT a

10 Convergence

Let fX1; X2; : : :g be a sequence of rv's and let X be another rv. Let Fn denote the cdf of Xn and let F denote the cdf of X.

Types of convergence

D

1. In distribution (weakly, in law): Xn ! X

lim Fn(t) = F (t) 8t where F continuous

n!1

P

2. In probability: Xn ! X

(8" > 0) lim P [jXn Xj > "] = 0

n!1

as

3. Almost surely (strongly): Xn ! X

P hn!1 Xn = Xi

=

P h

!

2

n!1 n

i

lim

 

 

: lim X (!) = X(!)

= 1

9

Соседние файлы в папке 1