概率论与数理统计
第二章
2.1 随机变量的概念
定义2.1: Ω \Omega Ω是样本空间, X = X ( ω ) X=X(\omega) X=X(ω)是该样本空间上的实值函数(定义域是样本空间), X X X称为随机变量,一般用 X , Y , Z , ξ , η , ς X,Y,Z,\xi,\eta,\varsigma X,Y,Z,ξ,η,ς表示
{ ω ∣ X ( ω ) = a } \{\omega|X(\omega)=a\} {ω∣X(ω)=a}事件: { X = a } \{X=a\} {X=a}事件
离散型:有限个/无限可列个
非离散型:主要研究连续型
2.2.1 离散型随机变量及其概率分布
X
X
X的所有取值
x
k
(
k
=
1
,
2
,
⋯
)
x_k(k=1,2,\cdots)
xk(k=1,2,⋯)(可列个)
P
(
X
=
x
k
)
=
P
k
P(X=x_k)=P_k
P(X=xk)=Pk 概率函数/概率分布
概率分布表:
X
1
0
P
1
2
1
2
\begin{array}{c|c|c} X & 1 & 0 \\ \hline P & \frac{1}{2} & \frac{1}{2} \\ \end{array}
XP121021
- P k ≥ 0 P_k\geq 0 Pk≥0
- ∑ P k = 1 \sum P_k=1 ∑Pk=1
连续型随机变量及其概率密度函数
- 每个小长方形的面积等于该组的频率
- 所有小长方形的面积之和等于1
- 介于 x = a x=a x=a和 x = b x=b x=b之间的面积近似等于 ( a , b ] (a,b] (a,b]之间的频率
定义:非负可积函数
f
(
x
)
f(x)
f(x),
f
(
x
)
≥
0
,
a
≤
b
f(x)\geq 0,a\leq b
f(x)≥0,a≤b
P
(
a
<
x
≤
b
)
=
∫
a
b
f
(
x
)
d
x
\displaystyle P(a<x\leq b)=\int_a^bf(x)dx
P(a<x≤b)=∫abf(x)dx
x
x
x:连续型随机变量
f
(
x
)
f(x)
f(x):
x
x
x的概率分布密度函数
记作
X
∼
f
(
x
)
X\sim f(x)
X∼f(x)
性质:
- f ( x ) ≥ 0 f(x)\geq 0 f(x)≥0
- ∫ − ∞ + ∞ f ( x ) = 1 \displaystyle\int_{-\infin}^{+\infin}f(x)=1 ∫−∞+∞f(x)=1
- 连续型随机变量取个别值的概率为0
连续型 不考虑端点
P
(
a
≤
x
≤
b
)
=
P
(
a
<
x
≤
b
)
=
P
(
a
≤
x
<
b
)
=
P
(
a
<
x
<
b
)
P(a\leq x\leq b)=P(a<x\leq b)=P(a\leq x<b)=P(a<x<b)
P(a≤x≤b)=P(a<x≤b)=P(a≤x<b)=P(a<x<b)
概率为0的事件未必是不可能事件
概率为1的事件未必是必然事件
X
X
X取
x
x
x附近值的概率大小
lim
Δ
x
→
0
P
(
x
<
X
<
x
+
Δ
x
)
Δ
x
=
∫
x
x
+
Δ
x
f
(
x
)
d
x
Δ
x
\displaystyle\lim\limits_{\Delta x\to 0} \frac{P(x<X<x+\Delta x)}{\Delta x}=\frac{\displaystyle\int_x^{x+\Delta x}f(x)dx}{\Delta x}
Δx→0limΔxP(x<X<x+Δx)=Δx∫xx+Δxf(x)dx
P ( x < X < x + Δ x ) ≈ f ( x ) Δ x P(x<X<x+\Delta x)\approx f(x)\Delta x P(x<X<x+Δx)≈f(x)Δx
2.2.2 分布函数的定义
定义:
F
(
x
)
=
P
(
X
≤
x
)
F(x)=P(X\leq x)
F(x)=P(X≤x)(普通的实函数)
X
X
X取值不超过
x
x
x的概率
x
∈
(
−
∞
,
+
∞
)
,
F
(
x
)
∈
[
0
,
1
]
x\in(-\infin,+\infin),F(x)\in[0,1]
x∈(−∞,+∞),F(x)∈[0,1]
离散型的分布函数
性质:
- 0 ≤ F ( x ) ≤ 1 , x ∈ ( − ∞ , + ∞ ) 0\leq F(x)\leq 1,x\in(-\infin,+\infin) 0≤F(x)≤1,x∈(−∞,+∞)
-
F
(
x
)
F(x)
F(x)不减:
∀
x
1
<
x
2
,
F
(
x
1
)
≤
F
(
x
2
)
\forall x_1<x_2,F(x_1)\leq F(x_2)
∀x1<x2,F(x1)≤F(x2)
lim x → + ∞ F ( x ) = F ( + ∞ ) = 1 \lim\limits_{x\to +\infin}F(x)=F(+\infin)=1 x→+∞limF(x)=F(+∞)=1
lim x → − ∞ F ( x ) = F ( − ∞ ) = 0 \lim\limits_{x\to -\infin}F(x)=F(-\infin)=0 x→−∞limF(x)=F(−∞)=0 -
F
(
x
)
F(x)
F(x)是右连续的,至多有可列个间断点
lim x → a + F ( x ) = F ( a ) \lim\limits_{x\to a^+}F(x)=F(a) x→a+limF(x)=F(a)
公式:
P
(
X
≤
a
)
=
F
(
a
)
P(X\leq a)=F(a)
P(X≤a)=F(a)
P
(
X
>
a
)
=
1
−
F
(
a
)
P(X>a)=1-F(a)
P(X>a)=1−F(a)
P
(
a
<
X
≤
b
)
=
P
(
X
≤
b
)
−
P
(
X
≤
a
)
=
F
(
b
)
−
F
(
a
)
P(a<X\leq b)=P(X\leq b)-P(X\leq a)=F(b)-F(a)
P(a<X≤b)=P(X≤b)−P(X≤a)=F(b)−F(a)
P
(
X
=
a
)
=
F
(
a
)
−
F
(
a
−
0
)
P(X=a)=F(a)-F(a-0)
P(X=a)=F(a)−F(a−0)
P
(
a
≤
X
≤
b
)
=
F
(
b
)
−
F
(
a
−
0
)
P(a\leq X\leq b)=F(b)-F(a-0)
P(a≤X≤b)=F(b)−F(a−0)
P
(
X
<
a
)
=
F
(
a
−
0
)
P(X<a)=F(a-0)
P(X<a)=F(a−0)
P
(
X
≥
a
)
=
1
−
F
(
a
−
0
)
P(X\geq a)=1-F(a-0)
P(X≥a)=1−F(a−0)
例题
【例1】
F
(
x
)
=
{
a
−
e
−
λ
x
x
>
0
0
x
≤
0
F(x)=\begin{cases} a-e^{-\lambda x} & x>0 \\ 0 & x\leq 0 \end{cases}
F(x)={a−e−λx0x>0x≤0
λ
>
0
\lambda>0
λ>0,求
a
a
a
解:
F
(
−
∞
)
=
0
=
0
F(-\infin)=0=0
F(−∞)=0=0
F
(
+
∞
)
=
lim
x
→
+
∞
(
a
−
1
e
λ
x
)
=
a
=
1
F(+\infin)=\displaystyle\lim\limits_{x\to+\infin}(a-\frac{1}{e^{\lambda x}})=a=1
F(+∞)=x→+∞lim(a−eλx1)=a=1
故
a
=
1
a=1
a=1
【例2】
离散型随机变量概率分布表如下:
X
−
1
2
3
P
1
2
1
3
1
6
\begin{array}{c|c} X & -1 & 2 & 3 \\ \hline P & \frac{1}{2} & \frac{1}{3} & \frac{1}{6} \\ \end{array}
XP−121231361
求分布函数
解:
F
(
x
)
=
P
(
X
≤
x
)
,
x
∈
(
−
∞
,
+
∞
)
F(x)=P(X\leq x), x\in(-\infin,+\infin)
F(x)=P(X≤x),x∈(−∞,+∞)
-
x
<
−
1
x<-1
x<−1
F ( x ) = P ( X ≤ x ) = 0 F(x)=P(X\leq x)=0 F(x)=P(X≤x)=0 -
−
1
≤
x
<
2
-1\leq x<2
−1≤x<2
F ( x ) = P ( X ≤ x ) = P ( X = − 1 ) = 1 / 2 F(x)=P(X\leq x)=P(X=-1)=1/2 F(x)=P(X≤x)=P(X=−1)=1/2 -
2
≤
x
<
3
2\leq x<3
2≤x<3
F ( x ) = P ( X ≤ x ) = P ( X = − 1 ) + P ( X = 2 ) = 5 / 6 F(x)=P(X\leq x)=P(X=-1)+P(X=2)=5/6 F(x)=P(X≤x)=P(X=−1)+P(X=2)=5/6 -
3
≤
x
3\leq x
3≤x
F ( x ) = P ( X ≤ x ) = P ( X = − 1 ) + P ( X = 2 ) + P ( X = 3 ) = 1 F(x)=P(X\leq x)=P(X=-1)+P(X=2)+P(X=3)=1 F(x)=P(X≤x)=P(X=−1)+P(X=2)+P(X=3)=1
F ( x ) = { 0 x < − 1 1 / 2 − 1 ≤ x < 2 5 / 6 2 ≤ x < 3 1 3 ≤ x F(x)=\begin{cases} 0 & x<-1 \\ 1/2 & -1\leq x<2 \\ 5/6 & 2\leq x<3 \\ 1 & 3\leq x \end{cases} F(x)=⎩⎪⎪⎪⎨⎪⎪⎪⎧01/25/61x<−1−1≤x<22≤x<33≤x
分布函数
→
\rarr
→概率函数
间断点
x
k
x_k
xk是
X
X
X的取值
P
(
X
=
x
k
)
=
F
(
x
k
)
−
F
(
x
k
−
0
)
P(X=x_k)=F(x_k)-F(x_k-0)
P(X=xk)=F(xk)−F(xk−0)
连续型的分布函数
F ( x ) = P ( X ≥ x ) = ∫ − ∞ x f ( t ) d t F(x)=P(X\geq x)=\displaystyle\int_{-\infin}^xf(t)dt F(x)=P(X≥x)=∫−∞xf(t)dt
例题
【例3】
f
(
x
)
=
1
π
(
1
+
x
2
)
\displaystyle f(x)=\frac{1}{\pi(1+x^2)}
f(x)=π(1+x2)1,求分布函数
解:
F
(
x
)
=
∫
−
∞
x
1
π
(
1
+
x
2
)
d
t
=
1
π
arctan
t
+
1
2
\displaystyle F(x)=\int_{-\infin}^x\frac{1}{\pi(1+x^2)}dt=\frac{1}{\pi}\arctan t +\frac{1}{2}
F(x)=∫−∞xπ(1+x2)1dt=π1arctant+21
【例4】
f
(
x
)
=
{
−
1
2
x
+
1
0
≤
x
≤
2
0
其
他
f(x)=\begin{cases} \displaystyle -\frac{1}{2}x+1 & 0\leq x\leq 2 \\ 0 & 其他 \end{cases}
f(x)=⎩⎨⎧−21x+100≤x≤2其他
解:
-
x
<
0
x<0
x<0
F ( x ) = ∫ − ∞ x f ( x ) d t = ∫ − ∞ x 0 d t = 0 \displaystyle F(x)=\int_{-\infin}^xf(x)dt=\int_{-\infin}^x0dt=0 F(x)=∫−∞xf(x)dt=∫−∞x0dt=0 -
0
≤
x
<
2
0\leq x<2
0≤x<2
F ( x ) = ∫ − ∞ x f ( x ) d t = ∫ − ∞ 0 0 d t + ∫ 0 x ( − 1 2 t + 1 ) = − 1 4 x 2 + x \displaystyle F(x)=\int_{-\infin}^xf(x)dt=\int_{-\infin}^00dt+\int_0^x(-\frac{1}{2}t+1)=-\frac{1}{4}x^2+x F(x)=∫−∞xf(x)dt=∫−∞00dt+∫0x(−21t+1)=−41x2+x -
2
≤
x
2\leq x
2≤x
F ( x ) = ∫ − ∞ x f ( x ) d t = ∫ − ∞ 0 0 d t + ∫ 0 2 ( − 1 2 t + 1 ) + ∫ 2 x 0 d t = 1 \displaystyle F(x)=\int_{-\infin}^xf(x)dt=\int_{-\infin}^00dt+\int_0^2(-\frac{1}{2}t+1)+\int_2^x0dt=1 F(x)=∫−∞xf(x)dt=∫−∞00dt+∫02(−21t+1)+∫2x0dt=1
【例5】
F
(
x
)
=
{
0
x
<
0
A
x
2
0
≤
x
<
1
1
1
≤
x
F(x)=\begin{cases} 0 & x<0 \\ Ax^2 & 0\leq x<1 \\ 1 & 1\leq x \end{cases}
F(x)=⎩⎪⎨⎪⎧0Ax21x<00≤x<11≤x,求:1.
A
A
A;2.
f
(
x
)
f(x)
f(x);3.
P
(
0.3
<
x
<
0.7
)
P(0.3<x<0.7)
P(0.3<x<0.7)
解:
-
求 A A A
F ( − ∞ ) = lim x → − ∞ F ( x ) = 0 F(-\infin)=\lim\limits_{x\to-\infin}F(x)=0 F(−∞)=x→−∞limF(x)=0
F ( + ∞ ) = lim x → + ∞ F ( x ) = 1 F(+\infin)=\lim\limits_{x\to+\infin}F(x)=1 F(+∞)=x→+∞limF(x)=1
lim x → 0 + F ( x ) = lim x → 0 + A x 2 = 0 = F ( 0 ) = 0 \lim\limits_{x\to0^+}F(x)=\lim\limits_{x\to0^+}Ax^2=0=F(0)=0 x→0+limF(x)=x→0+limAx2=0=F(0)=0
lim x → 1 − F ( x ) = lim x → 1 − A x 2 = A = F ( 1 ) = 1 \lim\limits_{x\to1^-}F(x)=\lim\limits_{x\to1^-}Ax^2=A=F(1)=1 x→1−limF(x)=x→1−limAx2=A=F(1)=1
故 A = 1 A=1 A=1 -
f ( x ) = { 2 x 0 ≤ x < 1 0 其 他 f(x)=\begin{cases} 2x & 0\leq x<1\\ 0 & 其他 \end{cases} f(x)={2x00≤x<1其他
-
P ( 0.3 < x < 0.7 ) = F ( 0.7 ) − F ( 0.3 ) = 0.49 − 0.09 = 0.4 P(0.3<x<0.7)=F(0.7)-F(0.3)=0.49-0.09=0.4 P(0.3<x<0.7)=F(0.7)−F(0.3)=0.49−0.09=0.4
P ( 0.3 < x < 0.7 ) = ∫ 0.3 0.7 f ( x ) d x = ∫ 0.3 0.7 2 x d x \displaystyle P(0.3<x<0.7)=\int_{0.3}^{0.7}f(x)dx=\int_{0.3}^{0.7}2xdx P(0.3<x<0.7)=∫0.30.7f(x)dx=∫0.30.72xdx
2.2.3 常见的分布
离散型常见分布
0-1分布
X 1 0 P p 1 − p \begin{array}{c|c} X & 1 & 0 \\ \hline P & p & 1-p \\ \end{array} XP1p01−p
P
(
X
=
k
)
=
p
k
(
1
−
p
)
1
−
k
P(X=k)=p^k(1-p)^{1-k}
P(X=k)=pk(1−p)1−k
(二项分布的特例)
- 有两种结果
- 试验只做一次
几何分布
P
(
A
)
=
p
P(A)=p
P(A)=p
第
k
k
k次首次发生,前
k
−
1
k-1
k−1次未发生
P
(
X
=
k
)
=
(
1
−
p
)
k
−
1
p
k
,
k
=
0
,
1
,
2
,
⋯
P(X=k)=(1-p)^{k-1}p^k,k=0,1,2,\cdots
P(X=k)=(1−p)k−1pk,k=0,1,2,⋯
X
∼
G
(
p
)
X\sim G(p)
X∼G(p)
二项分布
P
(
A
)
=
p
P(A)=p
P(A)=p
n
n
n次试验,发生了
k
k
k次
P
(
X
=
k
)
=
C
n
k
p
k
(
1
−
p
)
n
−
k
,
k
=
0
,
1
,
2
,
⋯
,
n
P(X=k)=C_n^kp^k(1-p)^{n-k}, k=0,1,2,\cdots,n
P(X=k)=Cnkpk(1−p)n−k,k=0,1,2,⋯,n
X
∼
B
(
n
,
p
)
X\sim B(n,p)
X∼B(n,p)
n = 1 n=1 n=1时, P ( X = k ) = C 1 k p k ( 1 − p ) 1 − k , k = 0 , 1 P(X=k)=C_1^kp^k(1-p)^{1-k},k=0,1 P(X=k)=C1kpk(1−p)1−k,k=0,1(0-1分布)
最可能值:
- ( n + 1 ) p (n+1)p (n+1)p不为整数, [ ( n + 1 ) p ] [(n+1)p] [(n+1)p]达到最大值
- ( n + 1 ) p (n+1)p (n+1)p为整数, ( n + 1 ) p (n+1)p (n+1)p和 ( n + 1 ) p + 1 (n+1)p+1 (n+1)p+1都是最大值
泊松分布
P
(
X
=
k
)
=
λ
k
k
!
e
−
λ
,
k
=
1
,
2
,
3
,
⋯
\displaystyle P(X=k)=\frac{\lambda^k}{k!}e^{-\lambda},k=1,2,3,\cdots
P(X=k)=k!λke−λ,k=1,2,3,⋯
λ
>
0
\lambda>0
λ>0
X
∼
P
(
λ
)
X\sim P(\lambda)
X∼P(λ)
电台收到的呼叫次数,公用设施(候车,收银台,一员挂号处)
计算方式:查表
二项分布可以用泊松分布近似
条件:
n
n
n较大,
p
p
p较小,
n
p
np
np适中(
n
≥
100
,
n
p
≤
10
n\geq100,np\leq10
n≥100,np≤10)
例题
【例5】电话台用户呼叫次数
X
∼
P
(
3
)
X\sim P(3)
X∼P(3),写出
X
X
X的概率函数,以及一分钟呼叫不超过5次的概率
解:
X
∼
P
(
3
)
X\sim P(3)
X∼P(3)
λ
=
3
\lambda=3
λ=3
P
(
X
=
k
)
=
λ
k
k
!
e
−
λ
=
3
k
k
!
e
−
3
,
k
=
0
,
1
,
2
,
⋯
\displaystyle P(X=k)=\frac{\lambda^k}{k!}e^{-\lambda}=\frac{3^k}{k!}e^{-3},k=0,1,2,\cdots
P(X=k)=k!λke−λ=k!3ke−3,k=0,1,2,⋯
P ( x ≤ 5 ) = ∑ k = 0 5 P ( X = k ) = 0.916 P(x\leq5)=\displaystyle\sum_{k=0}^5P(X=k)=0.916 P(x≤5)=k=0∑5P(X=k)=0.916
【例6】证券营业部有1000个账户,每户存了10万元,每个账户来提20%的概率时0.006,问应该准备多少现金,才能保证95%以上的概率满足提款要求
解:
10
×
0.2
=
2
10\times0.2=2
10×0.2=2(万元)
设:
X
X
X:提钱的用户数,
X
∼
B
(
1000
,
0.006
)
X\sim B(1000,0.006)
X∼B(1000,0.006),准备现金
x
x
x元
P
(
2
X
≤
x
)
≥
0.95
P(2X\leq x)\geq 0.95
P(2X≤x)≥0.95
P
(
X
≤
x
2
)
≥
0.95
\displaystyle P(X\leq \frac{x}{2})\geq 0.95
P(X≤2x)≥0.95
n
=
1000
,
p
=
0.006
,
n
p
=
6
n=1000,p=0.006,np=6
n=1000,p=0.006,np=6
⇒
λ
=
6
\Rarr\lambda=6
⇒λ=6
∑ k = 0 x 2 6 k k ! e − 6 ≥ 0.95 , x 2 ≥ 10 , x ≥ 20 \displaystyle\sum_{k=0}^{\frac{x}{2}}\frac{6^k}{k!}e^{-6} \geq0.95,\displaystyle\frac{x}{2}\geq10,x\geq20 k=0∑2xk!6ke−6≥0.95,2x≥10,x≥20
x = 20 x=20 x=20
【例7】
某种非传染病发病率为0.001,某单位有5000人,至少两人的病的概率?
解:
X
X
X:得病的人数,
X
∼
B
(
5000
,
0.001
)
X\sim B(5000,0.001)
X∼B(5000,0.001)
P
(
X
≥
2
)
=
1
−
P
(
X
=
0
)
−
P
(
X
=
1
)
P(X\geq2)=1-P(X=0)-P(X=1)
P(X≥2)=1−P(X=0)−P(X=1)
n
=
5000
,
p
=
0.001
,
n
p
=
5
n=5000,p=0.001,np=5
n=5000,p=0.001,np=5
⇒
λ
=
5
\Rarr\lambda=5
⇒λ=5
P
(
X
≥
2
)
=
1
−
0.006738
−
0.03369
=
0.959572
P(X\geq2)=1-0.006738-0.03369=0.959572
P(X≥2)=1−0.006738−0.03369=0.959572
超几何分布
100学生,男生60人,女生40人,任取10人,
X
X
X:取到的男生人数
P
(
X
=
k
)
=
C
60
k
C
40
10
−
k
C
100
10
\displaystyle P(X=k)=\frac{C_{60}^kC_{40}^{10-k}}{C_{100}^{10}}
P(X=k)=C10010C60kC4010−k
定义:
N
N
N个元素,
N
1
N_1
N1个属于第一类,
N
2
N_2
N2个属于第二类,取
n
n
n个,
X
X
X:
n
n
n个中属于第一类的个数
P
(
X
=
k
)
=
C
N
1
k
C
N
2
n
−
k
C
N
n
,
k
=
0
,
1
,
2
,
⋯
,
min
{
n
,
N
1
}
\displaystyle P(X=k)=\frac{C_{N_1}^kC_{N_2}^{n-k}}{C_N^n},k=0,1,2,\cdots,\min\{n,N_1\}
P(X=k)=CNnCN1kCN2n−k,k=0,1,2,⋯,min{n,N1}
超几何分布可以用来描述不放回抽样的实验
当
N
N
N很大,
n
n
n相对
N
N
N很小时,
p
=
M
N
p=\frac{M}{N}
p=NM改变甚微,不放回抽样可以看作放回抽样
P
(
X
=
k
)
=
C
M
k
C
N
−
M
n
−
k
C
N
n
≈
C
n
k
p
k
(
1
−
p
)
n
−
k
\displaystyle P(X=k)=\frac{C_M^kC_{N-M}^{n-k}}{C_N^n}\approx C_n^kp^k(1-p)^{n-k}
P(X=k)=CNnCMkCN−Mn−k≈Cnkpk(1−p)n−k
例题
【例9】
10000粒种子,发芽率99%,取200粒至多1粒不发芽的概率?
解:
10000
×
(
1
−
99
%
)
=
100
10000\times(1-99\%)=100
10000×(1−99%)=100(粒)
发芽9900粒,不发芽100粒
N
1
=
100
,
N
2
=
9900
,
n
=
200
N_1=100,N_2=9900,n=200
N1=100,N2=9900,n=200
P
(
X
≤
1
)
=
P
(
X
=
0
)
+
P
(
X
=
1
)
=
C
100
0
C
9900
200
C
10000
200
+
C
100
1
C
9900
199
C
10000
200
\displaystyle P(X\leq1)=P(X=0)+P(X=1)=\frac{C_{100}^0C_{9900}^{200}}{C_{10000}^{200}}+\frac{C_{100}^1C_{9900}^{199}}{C_{10000}^{200}}
P(X≤1)=P(X=0)+P(X=1)=C10000200C1000C9900200+C10000200C1001C9900199
用二项分布近似:
n
=
200
,
p
=
0.01
n=200,p=0.01
n=200,p=0.01
P
(
X
≤
1
)
=
C
200
0
0.0
1
0
0.9
9
200
+
C
200
1
0.0
1
1
0.9
9
199
P(X\leq1)=C_{200}^00.01^00.99^{200}+C_{200}^10.01^10.99^{199}
P(X≤1)=C20000.0100.99200+C20010.0110.99199
用泊松分布近似:
n
=
200
≥
100
n=200\geq 100
n=200≥100较大,
p
=
0.01
p=0.01
p=0.01较小,
n
p
=
2
≤
10
np=2\leq 10
np=2≤10大小适中
λ
=
2
\lambda=2
λ=2
k
=
1
,
0
k=1,0
k=1,0
查表:
P
=
0.1353
+
0.2707
=
0.406
P=0.1353+0.2707=0.406
P=0.1353+0.2707=0.406
连续型常见分布
均匀分布
f ( x ) = { 1 b − a a ≤ x ≤ b 0 e l s e f(x)=\begin{cases} \displaystyle\frac{1}{b-a} & a\leq x\leq b\\ 0 & else \end{cases} f(x)=⎩⎨⎧b−a10a≤x≤belse
X ∼ U [ a , b ] X\sim U[a,b] X∼U[a,b]
分布函数:
F
(
x
)
=
∫
−
∞
x
f
(
t
)
d
t
=
{
0
x
<
a
x
−
a
b
−
a
a
≤
x
<
b
1
x
≤
b
F(x)=\displaystyle\int_{-\infin}^xf(t)dt=\begin{cases} 0 & x<a \\ \displaystyle\frac{x-a}{b-a} & a\leq x<b \\ 1 & x\leq b \end{cases}
F(x)=∫−∞xf(t)dt=⎩⎪⎪⎨⎪⎪⎧0b−ax−a1x<aa≤x<bx≤b
X
∼
[
a
,
b
]
,
[
c
,
d
]
⊂
[
a
,
b
]
X\sim[a,b],[c,d]\subset[a,b]
X∼[a,b],[c,d]⊂[a,b]
P
(
c
≤
x
≤
d
)
=
∫
c
d
1
b
−
a
d
t
=
d
−
c
b
−
a
P(c\leq x\leq d)=\displaystyle\int_c^d\frac{1}{b-a}dt=\frac{d-c}{b-a}
P(c≤x≤d)=∫cdb−a1dt=b−ad−c
落在
[
a
,
b
]
[a,b]
[a,b]上任意子区间的概率与子区间的长度成正比,与子区间的位置无关
指数分布
f
(
x
)
=
{
λ
e
−
λ
x
x
>
0
0
x
≤
0
f(x)=\begin{cases} \lambda e^{-\lambda x} & x>0\\ 0 & x\leq0 \end{cases}
f(x)={λe−λx0x>0x≤0
λ
>
0
,
X
∼
E
x
p
(
λ
)
\lambda>0,X\sim E_{xp}(\lambda)
λ>0,X∼Exp(λ)
F ( x ) = { 1 − e − λ X x > 0 0 x ≤ 0 F(x)=\begin{cases} 1-e^{-\lambda X} & x>0\\ 0 & x\leq0 \end{cases} F(x)={1−e−λX0x>0x≤0
服务系统的服务时间,电话的通话时间,消耗性产品的寿命
正态分布
密度函数:
ϕ
(
x
)
=
1
2
π
σ
e
−
(
x
−
μ
)
2
2
σ
2
,
−
∞
<
x
<
+
∞
\phi(x)=\displaystyle\frac{1}{\sqrt{2\pi}\sigma}e^{-\displaystyle\frac{(x-\mu)^2}{2\sigma^2}},-\infin<x<+\infin
ϕ(x)=2πσ1e−2σ2(x−μ)2,−∞<x<+∞
记作 X ∼ N ( μ , σ 2 ) X\sim N(\mu,\sigma^2) X∼N(μ,σ2)
已知 ∫ − ∞ + ∞ e − x 2 d x \displaystyle\int_{-\infin}^{+\infin}e^{-x^2}dx ∫−∞+∞e−x2dx(高数知识)
则有:
∫
−
∞
+
∞
Φ
(
x
)
d
x
=
∫
−
∞
+
∞
1
2
π
σ
e
−
(
x
−
μ
)
2
2
σ
2
d
x
=
1
2
π
σ
∫
−
∞
+
∞
e
−
(
x
−
μ
)
2
2
σ
2
d
x
=
2
σ
2
π
σ
∫
−
∞
+
∞
e
−
(
x
−
μ
2
σ
)
2
d
(
x
−
μ
2
σ
)
=
1
π
π
=
1
\displaystyle\int_{-\infin}^{+\infin}\Phi(x)dx\\ =\displaystyle\int_{-\infin}^{+\infin}\displaystyle\frac{1}{\sqrt{2\pi}\sigma}e^{-\frac{(x-\mu)^2}{2\sigma^2}}dx\\ =\displaystyle\frac{1}{\sqrt{2\pi}\sigma}\displaystyle\int_{-\infin}^{+\infin}e^{-\frac{(x-\mu)^2}{2\sigma^2}}dx\\ =\displaystyle\frac{\sqrt{2}\sigma}{\sqrt{2\pi}\sigma}\displaystyle\int_{-\infin}^{+\infin}e^{-(\frac{x-\mu}{\sqrt{2}\sigma})^2}d(\frac{x-\mu}{\sqrt{2}\sigma})\\ =\frac{1}{\sqrt{\pi}}\sqrt{\pi}\\ =1
∫−∞+∞Φ(x)dx=∫−∞+∞2πσ1e−2σ2(x−μ)2dx=2πσ1∫−∞+∞e−2σ2(x−μ)2dx=2πσ2σ∫−∞+∞e−(2σx−μ)2d(2σx−μ)=π1π=1
分布函数:
Φ
(
x
)
=
1
2
π
σ
∫
−
∞
x
e
−
(
x
−
μ
)
2
2
σ
2
d
t
\Phi(x)=\displaystyle\frac{1}{\sqrt{2\pi}\sigma}\int_{-\infin}^xe^{-\frac{(x-\mu)^2}{2\sigma^2}}dt
Φ(x)=2πσ1∫−∞xe−2σ2(x−μ)2dt
性质:
- y = ϕ ( x ) y=\phi(x) y=ϕ(x)以 x = μ x=\mu x=μ为对称轴
- x = μ x=\mu x=μ时, ϕ ( x ) \phi(x) ϕ(x)取最大值 1 2 π σ \frac{1}{\sqrt{2\pi}\sigma} 2πσ1
- y = ϕ ( x ) y=\phi(x) y=ϕ(x)以 x x x轴为渐近线, x = μ ± σ x=\mu\pm\sigma x=μ±σ时有拐点
- σ \sigma σ固定, μ \mu μ变化:图像左右移动
- μ \mu μ固定, σ \sigma σ变化: σ \sigma σ变小,最高点上移(变陡); σ \sigma σ变大,最高点下移(变缓)
标准正态分布:
μ
=
0
,
σ
=
1
\mu=0,\sigma=1
μ=0,σ=1
ϕ
0
(
x
)
=
1
2
π
e
−
x
2
2
,
−
∞
<
x
<
+
∞
\phi_0(x)=\frac{1}{\sqrt{2\pi}}e^{-\frac{x^2}{2}},-\infin<x<+\infin
ϕ0(x)=2π1e−2x2,−∞<x<+∞
Φ
0
(
x
)
=
1
2
π
∫
−
∞
x
e
−
t
2
2
d
t
\Phi_0(x)=\frac{1}{\sqrt{2\pi}}\displaystyle\int_{-\infin}^xe^{-\frac{t^2}{2}}dt
Φ0(x)=2π1∫−∞xe−2t2dt
性质:
y
y
y轴为对称轴(偶函数)
ϕ
0
(
x
)
=
ϕ
0
(
−
x
)
\phi_0(x)=\phi_0(-x)
ϕ0(x)=ϕ0(−x)
Φ
0
(
−
x
)
=
1
−
Φ
0
(
x
)
\Phi_0(-x)=1-\Phi_0(x)
Φ0(−x)=1−Φ0(x)
如果一个指标的影响因素有很多,每个因素起的作用都不太大,则这个指标服从正态分布
计算:
{
0
≤
x
≤
5
查
表
x
≥
5
ϕ
0
(
x
)
=
0
,
Φ
0
(
x
)
=
1
x
≤
−
5
ϕ
0
(
x
)
=
0
,
Φ
0
(
x
)
=
0
−
5
≤
x
≤
0
ϕ
0
(
x
)
=
ϕ
0
(
−
x
)
,
Φ
0
(
−
x
)
=
1
−
Φ
0
(
x
)
\begin{cases} 0\leq x\leq5 & 查表\\ x\geq5 & \phi_0(x)=0,\Phi_0(x)=1\\ x\leq-5 & \phi_0(x)=0,\Phi_0(x)=0\\ -5\leq x\leq0 & \phi_0(x)=\phi_0(-x),\Phi_0(-x)=1-\Phi_0(x) \end{cases}
⎩⎪⎪⎪⎨⎪⎪⎪⎧0≤x≤5x≥5x≤−5−5≤x≤0查表ϕ0(x)=0,Φ0(x)=1ϕ0(x)=0,Φ0(x)=0ϕ0(x)=ϕ0(−x),Φ0(−x)=1−Φ0(x)
一般正态分布向标准正态分布转化:
ϕ
(
x
)
=
1
σ
ϕ
0
(
x
−
μ
σ
)
\phi(x)=\displaystyle\frac{1}{\sigma}\phi_0(\frac{x-\mu}{\sigma})
ϕ(x)=σ1ϕ0(σx−μ)
Φ
(
x
)
=
Φ
0
(
x
−
μ
σ
)
\Phi(x)=\displaystyle\Phi_0(\frac{x-\mu}{\sigma})
Φ(x)=Φ0(σx−μ)
X
∼
N
(
μ
,
σ
2
)
X\sim N(\mu,\sigma^2)
X∼N(μ,σ2)
P
(
∣
X
−
μ
∣
<
σ
)
=
0.6826
P(|X-\mu|<\sigma)=0.6826
P(∣X−μ∣<σ)=0.6826
P
(
∣
X
−
μ
∣
<
2
σ
)
=
0.9544
P(|X-\mu|<2\sigma)=0.9544
P(∣X−μ∣<2σ)=0.9544
P
(
∣
X
−
μ
∣
<
3
σ
)
=
0.9974
P(|X-\mu|<3\sigma)=0.9974
P(∣X−μ∣<3σ)=0.9974
3
σ
3\sigma
3σ准则
如果一个系统设计时服从正态分布,在检验时不符合
3
σ
3\sigma
3σ准则,则不合格
X
∼
(
0
,
1
)
X\sim(0,1)
X∼(0,1),给定
α
(
0
<
α
<
1
)
\alpha(0<\alpha<1)
α(0<α<1),找到
u
α
u_\alpha
uα满足
P
(
X
>
u
α
)
=
α
P(X>u_\alpha)=\alpha
P(X>uα)=α,
u
α
u_\alpha
uα称为上
α
\alpha
α分位数
u
0.05
=
1.645
u_{0.05}=1.645
u0.05=1.645
u
0.025
=
1.96
u_{0.025}=1.96
u0.025=1.96
u
0.01
=
2.33
u_{0.01}=2.33
u0.01=2.33
2.3.1 随机变量函数的分布
已知 X X X是某分布,求 Y = f ( X ) Y=f(X) Y=f(X)是什么分布
离散型
例
已知:
X
7
8
9
10
P
0.1
0.3
0.4
0.2
\begin{array}{c|c} X & 7 & 8 & 9 & 10 \\ \hline P & 0.1 & 0.3 & 0.4 & 0.2 \\ \end{array}
XP70.180.390.4100.2
Y
=
4
X
Y=4X
Y=4X
则有:
Y
28
32
36
40
P
0.1
0.3
0.4
0.2
\begin{array}{c|c} Y & 28 & 32 & 36 & 40 \\ \hline P & 0.1 & 0.3 & 0.4 & 0.2 \\ \end{array}
YP280.1320.3360.4400.2
连续型
设 X X X的密度函数是 f X ( x ) , y = g ( x ) , Y = g ( X ) f_X(x),y=g(x),Y=g(X) fX(x),y=g(x),Y=g(X)
- F Y ( x ) → F X ( x ) F_Y(x)\rarr F_X(x) FY(x)→FX(x)
- 两边求导: f Y ( x ) ← f X ( x ) f_Y(x)\larr f_X(x) fY(x)←fX(x)
例题
【例1】
已知:
X
X
X的密度函数为
f
X
(
x
)
f_X(x)
fX(x),
Y
=
3
X
+
2
Y=3X+2
Y=3X+2
解:
F
Y
(
x
)
=
P
(
Y
≤
x
)
=
P
(
3
X
+
2
≤
x
)
=
P
(
X
≤
x
−
2
3
)
=
F
X
(
x
−
2
3
)
F_Y(x)=P(Y\leq x)=P(3X+2\leq x)=P(X\leq\displaystyle\frac{x-2}{3})=F_X(\displaystyle\frac{x-2}{3})
FY(x)=P(Y≤x)=P(3X+2≤x)=P(X≤3x−2)=FX(3x−2)
两边求导,得:
f
Y
(
x
)
=
1
3
f
X
(
x
−
2
3
)
f_Y(x)=\displaystyle\frac{1}{3}f_X(\frac{x-2}{3})
fY(x)=31fX(3x−2)
特别地,若
f
X
(
x
)
=
{
1
4
0
≤
x
≤
4
0
e
l
s
e
f_X(x)=\begin{cases} \displaystyle\frac{1}{4} & 0\leq x\leq4\\ 0 & else \end{cases}
fX(x)=⎩⎨⎧4100≤x≤4else
则
f
Y
(
x
)
=
{
1
12
2
≤
x
≤
14
0
e
l
s
e
f_Y(x)=\begin{cases} \displaystyle\frac{1}{12} & 2\leq x\leq14\\ 0 & else \end{cases}
fY(x)=⎩⎨⎧12102≤x≤14else
若 X X X服从 [ a , b ] [a,b] [a,b]的均匀分布,则 Y = k X + c Y=kX+c Y=kX+c也服从相应区间( [ k a + c , k b + c ] [ka+c,kb+c] [ka+c,kb+c]或 [ k b + c , k a + c ] [kb+c,ka+c] [kb+c,ka+c])上的均匀分布。
【例2】
X
∼
N
(
μ
,
σ
2
)
,
Y
=
a
X
+
b
,
a
≠
0
X\sim N(\mu,\sigma^2),Y=aX+b,a\not=0
X∼N(μ,σ2),Y=aX+b,a=0
解:
-
a > 0 a>0 a>0时, F Y ( x ) = P ( Y ≤ x ) = P ( a X + b ≤ x ) = P ( X ≤ x − b a ) = Φ ( x − b a ) F_Y(x)=P(Y\leq x)=P(aX+b\leq x)=P(X\leq\displaystyle\frac{x-b}{a})=\Phi(\displaystyle\frac{x-b}{a}) FY(x)=P(Y≤x)=P(aX+b≤x)=P(X≤ax−b)=Φ(ax−b)
f Y ( x ) = 1 a ϕ ( x − b a ) = 1 2 π σ e − ( x − b a − μ ) 2 2 σ 2 = 1 2 π a σ e − ( x − ( b + a μ ) ) 2 2 a 2 σ 2 f_Y(x)=\displaystyle\frac{1}{a}\phi(\frac{x-b}{a})=\frac{1}{\sqrt{2\pi}\sigma}e^{-\displaystyle\frac{(\frac{x-b}{a}-\mu)^2}{2\sigma^2}}=\frac{1}{\sqrt{2\pi}a\sigma}e^{-\displaystyle\frac{(x-(b+a\mu))^2}{2a^2\sigma^2}} fY(x)=a1ϕ(ax−b)=2πσ1e−2σ2(ax−b−μ)2=2πaσ1e−2a2σ2(x−(b+aμ))2 -
a < 0 a<0 a<0时, f Y ( x ) = 1 2 π ( − a ) σ e − ( x − ( b + a μ ) ) 2 2 a 2 σ 2 f_Y(x)=\frac{1}{\sqrt{2\pi}(-a)\sigma}e^{-\displaystyle\frac{(x-(b+a\mu))^2}{2a^2\sigma^2}} fY(x)=2π(−a)σ1e−2a2σ2(x−(b+aμ))2
故有 Y ∼ N ( a μ + b , ( a σ ) 2 ) Y\sim N(a\mu+b,(a\sigma)^2) Y∼N(aμ+b,(aσ)2)
定理:
若
X
X
X的密度函数为
f
X
(
x
)
,
Y
=
k
X
+
b
(
k
≠
0
)
,
f
Y
(
x
)
=
1
∣
k
∣
f
X
(
x
−
b
k
)
f_X(x),Y=kX+b(k\not=0),f_Y(x)=\displaystyle\frac{1}{|k|}f_X(\frac{x-b}{k})
fX(x),Y=kX+b(k=0),fY(x)=∣k∣1fX(kx−b)
【例3】
X
∼
N
(
0
,
1
)
,
Y
=
X
2
X\sim N(0,1),Y=X^2
X∼N(0,1),Y=X2
解:
x
<
0
x<0
x<0时,
F
Y
(
x
)
=
P
(
Y
≤
x
)
=
P
(
X
2
≤
x
)
=
0
F_Y(x)=P(Y\leq x)=P(X^2\leq x)=0
FY(x)=P(Y≤x)=P(X2≤x)=0
x
≥
0
x\geq0
x≥0时,
F
Y
(
x
)
=
P
(
Y
≤
x
)
=
P
(
X
2
≤
x
)
=
P
(
−
x
≤
X
≤
x
)
=
∫
−
x
x
1
2
π
e
−
t
2
2
d
t
=
2
∫
0
x
1
2
π
e
−
t
2
2
d
t
F_Y(x)=P(Y\leq x)=P(X^2\leq x)=P(-\sqrt{x}\leq X\leq\sqrt{x})=\displaystyle\int_{-\sqrt{x}}^{\sqrt{x}}\frac{1}{\sqrt{2\pi}}e^{-\frac{t^2}{2}}dt=2\displaystyle\int_0^{\sqrt{x}}\frac{1}{\sqrt{2\pi}}e^{-\frac{t^2}{2}}dt
FY(x)=P(Y≤x)=P(X2≤x)=P(−x≤X≤x)=∫−xx2π1e−2t2dt=2∫0x2π1e−2t2dt
f
Y
(
x
)
=
2
1
2
π
e
−
x
2
1
2
x
−
1
2
=
1
2
π
e
−
x
2
x
−
1
2
f_Y(x)=\displaystyle2\frac{1}{\sqrt{2\pi}}e^{-\frac{x}{2}}\frac{1}{2}x^{-\frac{1}{2}}=\displaystyle\frac{1}{\sqrt{2\pi}}e^{-\frac{x}{2}}x^{-\frac{1}{2}}
fY(x)=22π1e−2x21x−21=2π1e−2xx−21
综上, f Y ( x ) = { 1 2 π e − x 2 x − 1 2 x > 0 0 x ≤ 0 f_Y(x)=\begin{cases} \displaystyle\frac{1}{\sqrt{2\pi}}e^{-\frac{x}{2}}x^{-\frac{1}{2}} & x>0\\ 0 & x\leq0 \end{cases} fY(x)=⎩⎨⎧2π1e−2xx−210x>0x≤0