0% found this document useful (0 votes)
31 views

Chapter 4 Continuous Probability Distribution

The document defines concepts related to continuous probability distributions, including distribution functions, probability density functions, expected value, and variance. It provides definitions, theorems, and examples for the exponential and gamma distributions. Key concepts covered are that the exponential distribution describes waiting times and the gamma distribution generalizes this to include the number of events before a threshold is reached. Formulas are given for the probability density functions, expected values, and variances of these distributions.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
31 views

Chapter 4 Continuous Probability Distribution

The document defines concepts related to continuous probability distributions, including distribution functions, probability density functions, expected value, and variance. It provides definitions, theorems, and examples for the exponential and gamma distributions. Key concepts covered are that the exponential distribution describes waiting times and the gamma distribution generalizes this to include the number of events before a threshold is reached. Formulas are given for the probability density functions, expected values, and variances of these distributions.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 53

Chapter 4

Continuous Probability Distribution


Definition 4.1
• Let Y denote any random variable. The (cumulative)
distribution function of Y, denote by, F(y) is such that
• F(y)=P(Y≤y), -∞ < y < ∞
Theorem 4.1

• Properties of Distribution Function 分配函數的性質:


• 1. F(y) is a non-decreasing function (非遞減) of y
• 2. F(∞) = 1 and F(−∞) = 0, i.e. 0 ≤ F(y) ≤ 1.
Definition 4.2
• Let Y denote a random variable with distribution function F(y),
Y is a continuous random variable if F(y) is continuous.
Example
1
• p(y)=𝐶𝑦2 pyqn-y, y=0,1,2, let p= , find F(y)
2
Definition 4.3

• Let F(y) be the distribution function for a continuous random


𝑑𝐹(𝑦)
variable, Y. Then f(y)= , wherever the derivative exists, is
𝑑𝑦
called the probability density function (p.d.f) for the random
variable Y.
Theorem 4.2:
• Properties of a Density Function for a continuous random
variable.
• 1. f(y)≥ 0 for all y −∞ ≤y≤ ∞

• 2. ‫׬‬−∞ 𝑓 𝑦 𝑑𝑦 = 1
Theorem 4.3
• If random variable Y has a density function f(y) and a < b , then
the probability that Y falls in the interval [a,b] is F(b)-F(a).
example
0, 𝑦<0
• F(y)=ቐ𝑦, 0 ≤ 𝑦 ≤ 1, find the p.d.f. of Y
1, 1< 𝑦
3𝑦 2 , 0 ≤ 𝑦 ≤ 1
• f(y)= ቊ , find F(y) and graph f(y) and F(y).
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑐𝑦, 0≤𝑦≤2
f(y)=ቊ , find c and P(1≤ 𝑦 ≤ 2).
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Exercise 4.11
𝑐𝑦 2 , 0≤𝑦≤2
• f(y)=ቊ , find c and F(y).
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Definition 4.5:
• Y is a continuous random variable with p.d.f f(y). The expected
value of Y is

E(Y)=‫׬‬−∞ 𝒚𝒇 𝒚 𝒅𝒚
Theorem 4.4
• Let g(y) be a function of Y, then the expected value of g(y) is

E(g(Y))=‫׬‬−∞ 𝒈(𝒚)𝒇 𝒚 𝒅𝒚
Theorem 4.5(3.3)
• Let Y be a continuous random variable with probability density
function f(y) and c be a constant, then E(c)=c
∞ ∞
• Proof: E(c)=‫׬‬−∞ 𝒄𝒇 𝒚 𝒅𝒚 = 𝒄 ‫׬‬−∞ 𝒇 𝒚 𝒅𝒚 = 𝑐.
Theorem 4.5 (3.4)
• Let Y be a continuous random variable with probability density
function f(y), g(Y) be a function of Y, and c be a constant. Then
E(cg(Y))=cE(g(Y))
∞ ∞
• Proof: E(cg(Y))= ‫׬‬−∞ 𝒄𝒈(𝒚)𝒇 𝒚 𝒅𝒚 = 𝒄 ‫׬‬−∞ 𝒈(𝒚)𝒇 𝒚 𝒅𝒚
= 𝒄E(g(Y))

Theorem 4.5 (3.5)
• Let Y be a continuous random variable with probability density
function f(y) and g1(Y), g2(Y),…, gk(Y) be k functions of Y and c1,
c2,…, ck are constants. Then
E(c1g1(Y)+ c2g2(Y)+… +ckgk(Y))=c1E(g1(Y))+c2E(g2(Y))+…+ckE(gk(Y))
Proof: E(c1g1(Y)+ c2g2(Y)+… +ckgk(Y))

= ‫׬‬−∞ (c1g1(Y)+ c2g2(Y)+… +ckgk(Y))𝒇 𝒚 𝒅𝒚
∞ ∞ ∞
=‫׬‬−∞ c1g1(Y)f y dy+ ‫׬‬−∞ c2g2(y)f y dy+…+ ‫׬‬−∞ ckgk(y)f y dy
= c1E(g1(Y))+c2E(g2(Y))+…+ckE(gk(Y))
Theorem 4.6
• Let Y be a continuous random variable with probability density
function f(y) and mean E(Y)= μ. Then
V(Y)=E((Y-μ)2)=E(Y2)-μ2
proof: E((Y-μ)2)=E(Y2-2μY+μ2)
=E(Y2)-2E(μY)+E(μ2)
=E(Y2)-2μE(Y)+E(μ2)
=E(Y2)-2μE(Y)+μ2
=E(Y2)-2μ2+μ2
=E(Y2)-μ2
3 2
𝑦 , 0≤𝑦≤2
• f(y)=ቐ8 , find E(Y) and V(y).
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
solution
3
• E(Y)=
2
3
• V(Y)=
20
1
(2 − 𝑦), 0≤𝑦≤2
• f(y)=൝2 , find E(Y) and V(y).
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
solution
2
• E(Y)=
3
2
• V(Y)=
9
0.2, −1 < 𝑦 < 0
• f(y)=ቐ0.2 + 1.2𝑦, 0 ≤ 𝑦 ≤ 1, find the p.d.f. of Y
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
solution
• E(Y)=0.4
41
• V(Y)=
150
Gamma Probability Distribution
• 從開店到第一個顧客上門經過多少時間
• (等到公車所需花的時間)
Exponential Distribution
• Waiting time W > 0 is a random variable
• F(𝜔)=P(W≤ 𝜔)= 1- P(W>𝜔)
• = 1-P(no chance in (0, ω) )
• = 1- 𝑒 −𝜆𝜔 0~ω的區間
• f(𝜔)= 𝜆𝑒 −𝜆𝜔 , let y= ω
−𝜆𝑦 1
• f(y)= 𝜆𝑒 , let 𝜆 = ,
𝜃
1 −𝑦
• f(y)= 𝑒 𝜃, y > 0.
𝜃
MGF of exponential distribution
∞ ty
• ty
M(t)=E(e )= e 𝒇
‫𝟎׬‬ 𝒚 𝒅𝒚
∞ ty 1 −𝑦
• = ‫ 𝟎׬‬e 𝜃 𝑒 𝜃 𝒅𝒚
∞ 1 (𝑡𝑦−𝑦)
• = ‫𝑒 𝜃 𝟎׬‬ 𝜃 𝒅𝒚

𝑦

∞ 1 𝑦(𝑡𝜃−1) ∞1 (
𝜃
)
• =‫𝒚𝒅 𝜃 𝑒 𝟎׬‬ = ‫𝑒 𝜃 𝟎׬‬ 1−𝑡𝜃 𝒅𝒚
𝜃
𝑦
− 𝜃
𝟏 ∞ 1 𝟏
• = ‫׬‬ 𝜃 𝑒 (
1−𝑡𝜃
)
𝒅𝒚 = = (1 − 𝑡𝜃)−1
𝟏−𝒕𝜽 𝟎 𝟏−𝒕𝜽
1−𝑡𝜃
• To prove that M’(t=0)=

• M’’(t=0)=
• Gamma Distribution G(𝛼, 𝛽)
• 等到第α班公車的時間,
• 𝛼 = 1 時 → exponentinal distribution 指數分配
• Y : waiting time , 第α次公車來的時間 假設α= 5
• F(y)=P(Y≤ 𝑦) = 1 − 𝑃(𝑌 ≥ 𝑦)這個時間內可能有0.1.2.3.4班公車來
𝑒 −𝜆𝑦 (𝜆𝑦)𝑘
• =1-σ𝛼−1
𝑘=0 → distribution function of Y
𝑘!

dF ( y)
f ( y) = = − ( −  )e −  y , 當 k = 0
dy
1 ( y)0 e− y ( y)1 (− )e− y
−[ + ] , 當k=1
1! 1!
• f(y)= 2  ( y)1  e− y ( y)2 (− )e− y
−[ + ] , 當k=2
2! 2!
3  (  y ) 2  e −  y (  y ) 3 ( − ) e −  y
−[ + ] , 當k=3
3! 3!

( − 1)  ( y) −2 e− y ( y) −1 (− )e− y
−[ + ] , 當 k =α- 1
( − 1)! ( −1)!
 y −1e− y 1
= , 令 =
( − 1 ) ! 
−y
 −1 
y e
= ~ y0
( − 1)! 
−y
 −1 
y e
( ) = ( − 1)! f ( y) = 
, 0 y
( )
−y
yα−1 e β
• f y = ൞ β𝛼Γ(α) , 0 ≤ y < ∞
0, otherwise

• 已知‫׬‬0 f y dy = 1
−y
∞ α−1 β
• →‫׬‬0 y e dy = βα ∙ Γ(α)
−y 1
−y[ −t]
tY ∞ ty yα−1 e β ∞yα−1 e β
• M t =E e = ‫׬‬0 e ∙ dy = ‫׬‬0 dy
βα Γ(α) βα Γ(α)
−y
1 ∞ α−1 ∗ β
• = α ‫׬‬ y e βΤ1−βt dy , 令β =
β Γ(α) 0 1−βt
1 1 β α 1 α
• = ∙ β∗α ∙ Γ α = ∙ =
βα ∙Γ α βα 1−βt 1−βt
−α
• = 1 − βt
• E Y = m′ t = 0 = −α 1 − βt −α−1
−β ቚt=0
−α−1
• = αβ 1 − βt ቚt=0 = αβ

• E Y 2 = m′′ t = 0 = αβ −α − 1 (1 − βt)−α−2 ቚt=0

• = αβ α + 1 β ∙ (1 − βt)−α−2 ቚt=0
• = αβ α + 1 β = αβ2 α + 1
• V Y = E Y 2 − [E Y ]2 = αβ2 α + 1 − α2 β2
• = α2 β2 + αβ2 − α2 β2 = αβ2
−y
yα−1 e β
• f y =ቐ βα Γ(α) ,0 ≤ y < ∞
0 , otherwise
−y
−y
∞ ∞ yα−1 e β ∞ α−1 β
• ‫׬‬0 f y dy = ‫׬‬0 dy = 1 ⇒ ‫׬‬0 y e dy = βα ∙ Γ(α)
βα Γ(α)

−y
−y
∞ yα−1 e β 1 ∞ α β
• E Y = ‫׬‬0 y ∙ βαΓ(α) dy = α ‫׬‬ y e dy
β Γ(α) 0
1 α+1
• = α ∙ β ∙ Γ(α + 1) = αβ
β Γ(α)
−y
−y
∞ 2 yα−1 e β 1 ∞ α+1 β
• E Y2 = ‫׬‬0 y ∙ dy = ‫׬‬ y e dy
βα Γ(α) α
β Γ(α) 0
1 α+2
• = ∙ β ∙ Γ(α + 2) = β2 α(α + 1)
βα Γ(α)
2
• V Y =E Y − [E Y ]2 = αβ2 α + 1 − α2 β2 = αβ2
Chi-Square distribution 卡方分配 𝜒 2

v
• 當α = , v=degrees of freedom 自由度 β = 2
2
• 標準常態分配
• Z 2 ~𝜒 2 (1)
• Z12 + Z22 ~𝜒 2 (2)
• Z12 + Z22 + ⋯ + Zn2 ~ 𝜒 2 (n)

2 v 2 2 v
• E 𝜒 = αβ = ∙ 2=v, V 𝜒 = αβ = ∙ 4 = 2v
2 2
Uniform Probability Distribution 均勻分配

• Definition 4.6:

1
, θ1 ≤ y ≤ θ2
If θ1 < θ2 , f y = ቐθ2−θ1
0 , otherwise
c , θ1 ≤ y ≤ θ2
• (p f):f y = ቊ
0 , otherwise
θ2 θ2 1
• ⇒ ‫׬‬θ c dy = 1 ⇒ cy ቚθ = 1 ⇒ c θ2 − θ1 = 1 ⇒ c =
1 1 θ2 −θ1

• *沒有mgf
θ2 1 1 θ2
• E Y = ‫׬‬θ y ∙ θ −θ dy = ∙ ‫׬‬θ y dy
1 2 1 θ2 −θ1 1

1 2 θ2
• = ∙ y ฬθ
2(θ2 −θ1 ) 1

θ22 −θ21 θ2 +θ1


• = =
2(θ2 −θ1 ) 2
θ2 2 1 1 1
• E Y 2
= ‫׬‬θ y ∙ dy = ∙ ∙ y 3 ฬθθ2
1 θ2 −θ1 θ2 −θ1 3 1
1
• = ∙ (θ32 − θ13 )
3(θ2 −θ1 )
1 θ22 +θ1 θ2 +θ21
• = ∙ (θ2 − θ1 )(θ22 + θ2 θ1 + θ12 ) =
3(θ2 −θ1 ) 3
θ22 +θ1 θ2 +θ21 θ22 +2θ1 θ2 +θ21
• V Y = E Y 2 − [E Y ]2 = −
3 4
1
• = [4θ22 + 4θ1 θ2 + 4θ12 − 3θ22 − 6θ1 θ2 − 3θ12 ]
12
1 (θ2 −θ1 )2
• = θ22 − 2θ1 θ2 + θ12 =
12 12
Exercise 4.38:

• Y has a uniform distribution over the interval ( 0 , 1 )


a. Find F y ?
b.Show that P a ≤ Y ≤ a + b , for a ≥ 0 , b ≥ 0 and a + b ≤ 1
depends only upon the value of b.
solution
1
= 1 ,0 < y < ∞
a. f y = ቐ1−0
0 , otherwise
y y
y
න f t dt = න 1 ∙ dt = t ቚ = y − 0 = y , 0 ≤ y ≤ ∞
F y =൞ 0 0 0
0 , y<0
a+b a+b
b. P a ≤ y ≤ a + b = ‫׬‬a 1 ∙ dy = yቚ a = a+b −a=b
• Normal Distribution 常態分配
• 鐘形對稱單峰
1 y−μ 2
• f y = exp − , −∞ < 𝑦 < ∞
2πσ2 2σ2
∞ ∞ 1 y−μ 2
• 已知‫׬‬−∞ f y dy = ‫׬‬−∞ exp − 𝑑𝑦 = 1
2πσ2 2σ2
∞ tY 1 y−μ 2
• M t =E etY = ‫׬‬−∞ e ∙ exp − 𝑑𝑦
2πσ2 2σ2
∞ 1 y2 −2yμ+μ2 −2σ2 ty
• = ‫׬‬−∞ exp − 𝑑𝑦
2πσ2 2σ2
∞ 1 y2 −2y μ+σ2 t +(μ+σ2 t)2 (μ+σ2 t)2 −μ2
• = ‫׬‬−∞ exp − + 𝑑𝑦
2πσ2 2σ2 2σ2
2
(μ+σ2 t)2 −μ2 ∞ 1 y−(μ+σ2 t)
• = exp ∙ ‫׬‬−∞ exp − 𝑑𝑦
2σ2 2πσ2 2σ2

2μσ2 t+σ4 t2 σ2 t2
• = exp = exp μt +
2σ2 2
2tσ2 σ2 t2
• E Y = m′ t = 0 = μ + exp μt + ቚ𝑡=0 = 𝜇
2 2

• E Y 2 = m′′ t = 0 = σ2 exp ቂμt +


σ2 t2 σ2 t2
ቃ + (μ + tσ2 )2 exp μt + ቚ𝑡=0
2 2
• = σ2 + μ2
• V Y = E Y 2 − [E Y ]2 = σ2 + μ2 − μ2 = σ2
2 Y−μ
• Y ~ N μ ,σ z= ⇒ 標準化
σ
Theorem 4.12
• Let Y be a random variable with f(y) , g(Y) is a function of Y.
tg(Y) ∞ tg(Y)
• Then the MGF of g(Y) is E e = ‫׬‬−∞ e ∙ f y dy

y−μ
• Let z = , z is the standard normal distribution
σ
MGF of standard normal distribution
∞ t∙y−μ 1 y−μ 2
• Mz t = E et𝑧 = ‫׬‬−∞ e σ ∙ exp − 𝑑𝑦
2πσ2 2σ2
∞ 1 y2 −2μy+μ2 −2tyσ+2tμσ
• = ‫׬‬−∞ 2
exp − 2 𝑑𝑦 =
2πσ 2σ
∞ 1 y2 −2y μ+tσ +(μ+tσ)2 (μ+tσ)2 −μ2 −2tμσ
‫׬‬−∞ 2πσ2 exp − 2σ2 +
2σ2
𝑑𝑦
t2
• = exp( )
2
2
• 當μ = 0 , σ = 1時
1 z2
• f z = ∙ exp(− )
2π 2

You might also like