You are on page 1of 3

STOR 435 Formula Sheet

1. P (U k) = P (X1 k, . . . , Xn k) = P (X1 k)
K k+1 n
P (Xn k) = ((
) )
K
2. P (U = k) = P (U k) P (U k + 1)

Basics
Event relationships
1. A = B A B B A
2. A B = mutually exclusive

Maximum
Let V = M ax{X1 , . . . , Xn }

Probability definition

1. P (V k) = P (X1 k, . . . , Xn k) = P (X1 k)
k
P (Xn k) = (( )n )
K

1. 0 P (A) 1, A
2. P () = 1P
n
3. P (A) = i=1 P (Ai ) if Ai , . . . , An partitions A

2. P (V = k) = P (V k) P (V k 1)
3. E(V ) = kk=1 P (V k) = K
k=1 [1 P (V < k)] = K
K
K
k=1 P (X1 < k) P (Xn < k) = K k=1

Probability rules
(
1. P (A|B) =

P (AB)
P (B)

if P (B) > 0

Calculus

0
if P (B) = 0
2. P (A B)
=
P
(A)P
(B|A)
Pn
3. P (B) = i=1 P (Ai )P (B|Ai )

Bernoulli trials
 
 
n k nk
n
P (X = k) =
p q
if X Bin(n, p) where
=
k
k

n!
(nk)!k!

R
R
1. udv = iv vdu
2. For (x, y) (u, v),
use Jacobian determinant.
x x

(x, y) u v
=

(u, v) y y


u v

|J| =

Distributions
Independence definition
1.
2.
3.

P (A B) = P (A)P (B)
P (A|B) = P (A)
P B|A) = P (B)
Disjoint blocks of independent events are independent
Mutually exclusive ; independent
Mutually independent pairwise independent (:)

Density curve (continuous RV)


1. fR (x) 0

2. f (x) dx = 1
3. P (a X b) =

Rb
A

f (u) du

Multivariate
Joint: P(X,Y ) (x, y) = P (X = x, Y = y)
Marginal: PX (x) = P (X = x) = y P(X,Y ) (x, y) = P (X =
x, Y = y)
If X = Y X and Y have the same distribution

Maximum and Minimum


Minimum
Let U = M in{X1 , . . . , Xn }

General facts
1. Same distribution if P (X = v) = P (Y = v).
P (a X b) = P (a Y b).
2. Equal if P (X = Y ) = 1.

Implies

Binomial: X Bin(n, p)
1. X = IA1 + + IAn . Var(X) = IA1 + + IAn =
n(E(IA1 ) [E(IA1 )]2 ) = npq
2. A and B are independent IA + IB
3. A B IA IB = IA

Binomial approximation
z
P (a X b) P ( a.5

= npq. Works if 3.

b+.5
)

where = np and

Exponential: X exp()
x
1. f (x) = e
(
1 ex where x 0
2. F (X) =
0 where x < 0
1
1
3. E(X) = , V ar(X) = 2

4. Survival: P (X > x) = 1 F (X) = ex


5. Memoryless property: P (X > x + t|X > x) = P (X > t)

Gamma: Sn Gamma(n, p)

CDF

Continuous analog of Poisson


1. Let Sn = X1 + .. + Xn where Xi exp()
ex n xn1
2. fSn =
where (n) = (n 1)!
(n)
n
n
3. E(Sn ) = , Var(Sn ) = 2

Continuous and step-wise


1. F (x) = P (X x), limx = 1
2. Step function: F (x) = u:ux P (X = x) P (X = x) =
F (x) F (X ) where F (X is the left limit
(a) P (a < X < b) = F (b ) F (a)
(b) P (a X < b) = F (b ) F (a )
(c) P (a < X b) = F (b) F (a)
(d) P (a X b) =RF (b) F (a )
x
d
F (x)
3. Continuous: F (x) = f (u)du f (x) = dx

Geometric: X Geom(p)
1. P (X = k) = q k1 p, P (X k) = q k1
k1
2. E(X) =
= =
k=1 P (X k) = q
q
3. Var(X) = p2

1
p

Hypergeometric X HG(n; N, B)

Joint, marginal, and independence

Sample n balls without


N =B+R
 replacement,

B R
1. P (X = b) =

r
N
n

B
2. E(X) = P (Ai ) = np where p =
n
N n
R
3. V ar(X) = (
)npq where q =
N 1
n

Independent Normal RV
1. X = z + X N (, 2 )
2. X1 , . . . , Xn iid for Xi N (i , i2 ) Sn := X1 + + Xn
N (1 + + n , 12 + + n2 )

Multinomial
Let X1 , . . . , Xn be iid with P (X1 = m) = pm .
1. Nm = IX1 =m + + IXn =m
n!
pn1 . . . pnmm
2. Then, P (N1 = n1 , . . . , Nm = nm ) = n1 !...n
m! 1
3. Cov(Nk , Nm ) = npk pm

Negative Binomial: X NegBin(n, p)


Number of failures until r successes
k r
k1
1. P (X = k) = k+r1
r1 q p , P (X k) = q
rq
2
2. E(X) = = p , V ar(X) = rq/p

Conditional joint and marginal


R
1. fY (y) = fy (y|X = x)dx
2. Multiplication rule: f (x, y) = fx (x)fy (y|X = x)
f (x, y)
3. fy (y|X = x) =
fx (x)
R
4. P (Y B|X
=
x)
= B fY (y|X = x)dy
R
5. P (B) = R P (B|X = x)fX (x)dx
6. E(Y ) = E(Y |X = x)fX (x)dx

Expectation, Variance, and Correlation


Expectation

Normal: X N(, 2 )
Continuous analog of binomial
1
1. f (x) = 2
exp( 21 2 (x )2 )
2
2. cdf: (z)

Poisson: X Poisson()
k

1. P (X = k) exp k!

1. RJoint CDF:
R xn F (x1 , . . . , xn ) = P (X1 x1 , . . . , X1 xn ) =
x1
.
.
.
= f (x1 , . . . , xn )dx1 . . . dxn

R
2. 1D Marginal: fxi (xi ) = Rn1 f (u1 , . . . , ui1 , xi , ui+1 , . . . , un
du1 . . . dui1 dui+1 . . . dun
3. Independence: f (x1 , . . . , xn ) = fx1 (x1 ) . . . fxn (xn )
F (x1 , . . . , xn ) = Fx1 (x1 ) . . . Fxn (xn ).
Also, f (x, y) =
fx (x)fy (y)
R
4. E(g(x1 , . . . , xnR)) = Rn g(x1 , . . . , xn )f (x1 , . . . , xn )dx1 . . . dxn .
Also, E(X) = xfx (x)dx
5. P (X dx, Y dy = f (x, y)dxdy

k1



2. cdf: E(X) =
k=0 ke
k! = k=1 e
(k1)!

Let X be a random variable and c be a constant


1. = xP
( (x), E(X) = x xP (X = x)
0 if A does not occurs
2. IA =
1 if A occurs
3. E(IA ) = P (A)
4. E(X + Y ) = E(X) + E(Y )
5. E(c) = c, E(cX) = cE(X)
6. If X and Y are independent, E(XY ) = E(X)E(Y )
= V ar(X) =
7. Let pj = P (Xi = j). E(x) = p1 +2p2 + +npn = (p1 + +
pn ) + (p2 + + pn ) + . . . (pn1 + pn ) + pn = ni=1 P (x j)

Uniform: X Uniform[a, b]
1
ba
= a+b
2 ,

Conditional Expectation

1. f (x) =

2. E(X)
V ar(X) = (ba)
12
ol(A)
3. P ((x1 , . . . , xn ) A) = VV ol(D)
A D
4. D = [a, b][c, d](x, y) is uniformly distributed over D X
U nif orm[a, b] and Y U nif orm[a, b]

Conditional expectation < Unconditional expectation


1. E(Y |X = x) = y yP (Y = y|X = x)
2. E(g(x, y)|X = x) = y g(x, y)P (Y = y|X = x)
3. E(Y ) = E[E(Y |X)] = x E(Y |X = x)P (X = x)

Variance

Correlation
Cov(X, Y )
=
SD(X)SD(Y )
2. 1 Corr(X, Y ) 1
Y
X
and X = Y
(a) X = X
X
Y
(b) E(X ) = E(Y ) = 0
(c) V ar(X ) = V ar(Y ) = 1
3. Corr(x, y) = 1 y = ax + b for a, b R

1. Corr(X, Y ) =
1.
2.
3.
4.
5.
6.
7.
8.

Cov(X, Y ) = E((X X )(Y Y )) = E(XY ) E(X)E(Y )


V ar(X) = E((X )2 ) = E(X 2 ) 2
Cov(X, X)=Var(X)
Cov(X, Y ) = Cov(Y, X)
Cov(IA , IB ) = P (A B) P (B)P (A))
V ar(X + Y ) = V ar(X) + V ar(Y ) + 2Cov(X, Y )
Cov(ai Xi , bj Yj ) = ai bi Cov(Xi , Yj )
X and Y are independent Corr(X, Y ) = 0 (:)

You might also like