概率公式
P
(
B
−
A
)
=
P(B-A)=
P(B−A)=
P
(
B
−
B
A
)
=
P
(
B
)
−
P
(
B
A
)
P(B-BA)=P(B)-P(BA)
P(B−BA)=P(B)−P(BA)
P
(
A
⋃
B
)
=
P
(
A
)
+
P
(
B
)
−
P
(
A
B
)
P(A\bigcup B)=P(A)+P(B)-P(AB)
P(A⋃B)=P(A)+P(B)−P(AB)
条件概率 乘法定理
P
(
B
∣
A
)
=
P
(
A
B
)
P
(
A
)
P(B|A)=\frac{P(AB)}{P(A)}
P(B∣A)=P(A)P(AB)
P
(
B
ˉ
∣
A
)
=
1
−
P
(
B
∣
A
)
P(\bar{B}|A)=1-P(B|A)
P(Bˉ∣A)=1−P(B∣A)
若 P ( A ) > 0 , 则 P ( A B ) = P ( A ) P ( B ∣ A ) 若P(A)>0,则P(AB)=P(A)P(B|A) 若P(A)>0,则P(AB)=P(A)P(B∣A)
若 P ( B ) > 0 , 则 P ( A B ) = P ( B ) P ( A ∣ B ) 若P(B)>0,则P(AB)=P(B)P(A|B) 若P(B)>0,则P(AB)=P(B)P(A∣B)
贝叶斯公式
P ( B k ∣ A ) = P ( B k ) P ( A ∣ B k ) ∑ i = 1 n P ( B i ) P ( A ∣ B i ) P({{B}_{k}}|A)=\frac{P({{B}_{k}})P(A|{{B}_{k}})}{\sum\limits_{i=1}^{n}{P({{B}_{i}})P(A|{{B}_{i}})}} P(Bk∣A)=i=1∑nP(Bi)P(A∣Bi)P(Bk)P(A∣Bk)
全概率公式
P ( A ) = ∑ k = 1 ∞ P ( B k ) P ( A ∣ B k ) P(A)=\sum\limits_{k=1}^{\infty }{P({{B}_{k}})P(A|{{B}_{k}})} P(A)=k=1∑∞P(Bk)P(A∣Bk)
常见概率分布
分布 | 数学标记 | 分布律 | 期望 | 方差 |
---|---|---|---|---|
二项分布 | B(𝑛,𝑝) | P ( X = k ) = C n k p k ( 1 − p ) n − k P(X=k)=C_n^kp^k(1-p)^{n-k} P(X=k)=Cnkpk(1−p)n−k | np | np(1-p) |
泊松分布 | π ( λ ) \pi(\lambda) π(λ) | P ( X = k ) = λ k e − λ k ! P(X=k)=\tfrac{\lambda^ke^{-\lambda}}{k!} P(X=k)=k!λke−λ | λ = n p \lambda=np λ=np | λ = n p \lambda=np λ=np |
分布 | 数学标记 | 分布函数 | 概率密度 | 期望 | 方差 | 特点 |
---|---|---|---|---|---|---|
均匀分布 | U(a,b) | F ( x ) = { 0 , x < a x − a b − a , a ≤ x ≤ b 1 , b < x F(x)=\begin{cases}0,x<a \\\frac{x-a}{b-a},a\le x\le b \\1,b<x\end{cases} F(x)=⎩⎪⎨⎪⎧0,x<ab−ax−a,a≤x≤b1,b<x | f ( x ) = { 1 b − a , a < x < b 0 , 其 他 f(x)=\begin{cases}\frac{1}{b-a},a<x<b \\0,其他\end{cases} f(x)={b−a1,a<x<b0,其他 | a + b 2 \frac{a+b}{2} 2a+b | ( a − b ) 2 12 \frac{(a-b)^2}{12} 12(a−b)2 | |
指数分布 | Γ ( 1 , θ ) \Gamma(1,\theta) Γ(1,θ) | F ( x ) = { 1 − e − θ x , x > 0 0 , x ≤ 0 F\left( x \right)=\left\{ \begin{matrix} 1-{{e}^{-\theta \ x}}, & x>0 \\ 0, & x\le 0 \\\end{matrix} \right. F(x)={1−e−θ x,0,x>0x≤0 | f ( x ) = { θ e − θ x , x > 0 0 , 其 他 f(x)=\left\{ \begin{matrix}\theta {{e}^{-\theta x}}, x>0 \\0,其他\\\end{matrix} \right. f(x)={θe−θx,x>00,其他 | 1 θ \frac{1}{\theta} θ1 | 1 θ 2 \frac{1}{{{\theta }^{2}}} θ21 | P { X > s + t ∣ X > s } = P { X > t } P\{X>s+t\mid X>s\}=P\{X>t\} P{X>s+t∣X>s}=P{X>t} |
正态分布 | N( μ , σ 2 \mu,\sigma^2 μ,σ2) | f(x)= 1 2 π σ e − ( x − μ ) 2 2 σ 2 \frac{1}{\sqrt{2\pi }\sigma} e^{-\frac{(x-\mu)^2}{2\sigma^2} } 2πσ1e−2σ2(x−μ)2 | μ \mu μ | σ 2 \sigma^2 σ2 | 标 准 化 ξ ∗ = ξ − μ σ 标准化\xi ^*=\frac{\xi-\mu}{\sigma } 标准化ξ∗=σξ−μ | |
标准正态分布 | N(0,1) | ϕ ( x ) = 1 2 π e − x 2 2 \phi (x)=\frac{1}{\sqrt{2\pi }}{{e}^{-\frac{{{x}^{2}}}{2}}} ϕ(x)=2π1e−2x2 | 0 | 1 | Φ ( − x ) = 1 − Φ ( x ) , Φ ( 0 ) = 0.5 \Phi (-x)=1-\Phi (x),\Phi (0)=0.5 Φ(−x)=1−Φ(x),Φ(0)=0.5 |
联合概率密度
联合概率密度性质
∫ − ∞ + ∞ ∫ − ∞ + ∞ f ( x , y ) d x d y = 1 \int_{-\infty }^{+\infty }{\int_{-\infty }^{+\infty }{f(x,y)dxdy}}=1 ∫−∞+∞∫−∞+∞f(x,y)dxdy=1
边缘概率密度
f
X
(
x
)
=
∫
−
∞
+
∞
f
(
x
,
y
)
d
y
{{f}_{X}}(x)=\int_{-\infty }^{+\infty }{f(x,y)dy}
fX(x)=∫−∞+∞f(x,y)dy
f
Y
(
y
)
=
∫
−
∞
+
∞
f
(
x
,
y
)
d
x
{{f}_{Y}}(y)=\int_{-\infty }^{+\infty }{f(x,y)dx}
fY(y)=∫−∞+∞f(x,y)dx
相互独立时有
f
(
x
,
y
)
=
f
X
(
x
)
f
Y
(
y
)
f(x,y)={{f}_{X}}(x){{f}_{Y}}(y)
f(x,y)=fX(x)fY(y)
期望
定义
∫ − ∞ + ∞ ∣ x f ( x ) ∣ d x \int_{-\infty }^{+\infty }{|xf(x)|dx} ∫−∞+∞∣xf(x)∣dx收敛则有期望
计算
E
X
=
∫
−
∞
+
∞
x
f
(
x
)
d
x
EX=\int_{-\infty }^{+\infty }{xf(x)dx}
EX=∫−∞+∞xf(x)dx
E
(
X
)
=
∫
−
∞
+
∞
∫
−
∞
+
∞
x
f
(
x
,
y
)
d
x
d
y
E(X)=\int_{-\infty }^{+\infty }{\int_{-\infty }^{+\infty }{xf(x,y)\text{d}x\text{d}y}}
E(X)=∫−∞+∞∫−∞+∞xf(x,y)dxdy
性质
E
(
a
X
+
b
Y
+
c
)
=
a
E
X
+
b
E
Y
+
c
E(aX+bY+c)=aEX+bEY+c
E(aX+bY+c)=aEX+bEY+c
X,Y相互独立时
E
(
X
Y
)
=
E
(
X
)
E
(
Y
)
.
E(XY)=E(X)E(Y).
E(XY)=E(X)E(Y).
已知X的分布,求 Y = g ( X ) Y=g(X) Y=g(X)期望
离散型
E
Y
=
E
[
g
(
X
)
]
=
∑
i
=
1
∞
g
(
x
i
)
p
i
.
EY=E[g(X)]=\sum\limits_{i=1}^{\infty }{g({{x}_{i}}){{p}_{i}}}.
EY=E[g(X)]=i=1∑∞g(xi)pi.
E
Z
=
E
[
g
(
X
,
Y
)
]
=
∑
j
=
1
∞
∑
i
=
1
∞
g
(
x
i
,
y
j
)
p
i
j
EZ=E[g(X,Y)]=\sum\limits_{j=1}^{\infty }{\sum\limits_{i=1}^{\infty }{g}({{x}_{i}},{{y}_{j}}){{p}_{ij}}}
EZ=E[g(X,Y)]=j=1∑∞i=1∑∞g(xi,yj)pij
连续型
E
Y
=
E
[
g
(
X
)
]
=
∫
−
∞
+
∞
g
(
x
)
f
(
x
)
d
x
.
EY=E[g(X)]=\int_{-\infty }^{+\infty }{g(x)f(x)dx}.
EY=E[g(X)]=∫−∞+∞g(x)f(x)dx.
E
Y
=
E
[
g
(
X
,
Y
)
]
=
∫
−
∞
+
∞
∫
−
∞
+
∞
g
(
x
,
y
)
f
(
x
,
y
)
d
x
d
y
EY=E[g(X,Y)]=\int_{-\infty }^{+\infty }{\int_{-\infty }^{+\infty }{g}(x,y)f(x,y)dxdy}
EY=E[g(X,Y)]=∫−∞+∞∫−∞+∞g(x,y)f(x,y)dxdy
方差
定义
D
X
=
E
[
(
X
−
E
X
)
2
]
DX=E[{{(X-EX)}^{2}}]
DX=E[(X−EX)2]
(离散型)
=
∑
i
=
1
∞
(
x
i
−
E
X
)
2
p
i
=\sum\limits_{i=1}^{\infty }{{{({{x}_{i}}-EX)}^{2}}{{p}_{i}}}
=i=1∑∞(xi−EX)2pi
(连续型)
=
∫
−
∞
+
∞
(
x
−
E
X
)
2
f
(
x
)
d
x
=\int_{-\infty }^{+\infty }{{{(x-EX)}^{2}}f(x)dx}
=∫−∞+∞(x−EX)2f(x)dx
计算方法
D ( X ) = E ( X 2 ) − [ E ( X ) ] 2 D(X)=E({{X}^{2}})-{{[E(X)]}^{2}} D(X)=E(X2)−[E(X)]2
性质
D
(
C
)
=
0
D(C)=0
D(C)=0
D
(
C
X
)
=
C
2
D
(
X
)
D(CX)={{C}^{2}}D(X)
D(CX)=C2D(X)
D
(
X
±
Y
)
=
D
(
X
)
+
D
(
Y
)
±
2
E
[
(
X
−
E
X
)
(
Y
−
E
Y
)
]
D(X\pm Y)=D(X)+D(Y)\pm 2E[(X-EX)(Y-EY)]
D(X±Y)=D(X)+D(Y)±2E[(X−EX)(Y−EY)]
=
D
(
X
)
+
D
(
Y
)
±
2
cov
(
X
,
Y
)
=D(X)+D(Y) \pm 2\operatorname{cov}(X,Y)
=D(X)+D(Y)±2cov(X,Y)
协方差
cov
(
X
,
Y
)
=
E
[
(
X
−
E
X
)
(
Y
−
E
Y
)
]
\operatorname{cov}(X,Y)=E[(X-EX)(Y-EY)]
cov(X,Y)=E[(X−EX)(Y−EY)]
(离散型)
=
∑
i
∑
j
(
x
i
−
E
X
)
(
y
j
−
E
Y
)
p
i
j
=\sum\limits_{i}{\sum\limits_{j}{({{x}_{i}}-EX)({{y}_{j}}-EY){{p}_{ij}}}}
=i∑j∑(xi−EX)(yj−EY)pij
(连续型)
=
∫
−
∞
+
∞
∫
−
∞
+
∞
(
x
−
E
X
)
(
y
−
E
Y
)
f
(
x
,
y
)
d
x
d
y
=\int_{-\infty }^{+\infty }{\int_{-\infty }^{+\infty }{(x-EX)(y-EY)f(x,y)dxdy}}
=∫−∞+∞∫−∞+∞(x−EX)(y−EY)f(x,y)dxdy
计算方法
cov
(
X
,
Y
)
=
E
(
X
Y
)
−
E
(
X
)
⋅
E
(
Y
)
\operatorname{cov}(X,Y)=E(XY)-E(X)\cdot E(Y)
cov(X,Y)=E(XY)−E(X)⋅E(Y)
协方差性质
cov
(
X
,
Y
)
=
cov
(
Y
,
X
)
\operatorname{cov}(X,Y)=\operatorname{cov}(Y,X)
cov(X,Y)=cov(Y,X)
D
X
=
cov
(
X
,
X
)
DX=\operatorname{cov}(X,X)
DX=cov(X,X)
cov
(
a
X
,
b
Y
)
=
a
b
cov
(
X
,
Y
)
\operatorname{cov}(aX,bY)=ab\operatorname{cov}(X,Y)
cov(aX,bY)=abcov(X,Y)
cov
(
X
1
+
X
2
,
Y
)
=
cov
(
X
1
,
Y
)
+
cov
(
X
2
,
Y
)
\operatorname{cov}({{X}_{1}}+{{X}_{2}},Y)=\operatorname{cov}({{X}_{1}},Y)+\operatorname{cov}({{X}_{2}},Y)
cov(X1+X2,Y)=cov(X1,Y)+cov(X2,Y)
cov
(
∑
i
=
1
n
a
i
X
i
,
∑
j
=
1
k
b
j
Y
j
)
=
\operatorname{cov}(\sum\limits_{i=1}^{n}{{{a}_{i}}{{X}_{i}}},\sum\limits_{j=1}^{k}{{{b}_{j}}{{Y}_{j}}})=
cov(i=1∑naiXi,j=1∑kbjYj)=
∑
i
=
1
n
∑
j
=
1
k
a
i
b
j
cov
(
X
i
,
Y
j
)
\sum\limits_{i=1}^{n}{\sum\limits_{j=1}^{k}{{{a}_{i}}{{b}_{j}}\operatorname{cov}({{X}_{i}},{{Y}_{j}})}}
i=1∑nj=1∑kaibjcov(Xi,Yj)
X
,
Y
相
互
独
立
,
则
有
cov
(
X
,
Y
)
=
0.
X,Y相互独立,则有\operatorname{cov}(X,Y)=0.
X,Y相互独立,则有cov(X,Y)=0.
例题
cov
(
X
1
+
X
2
,
Y
)
=
cov
(
X
1
,
Y
)
+
cov
(
X
2
,
Y
)
\operatorname{cov}({{X}_{1}}+{{X}_{2}},Y)=\operatorname{cov}({{X}_{1}},Y)+\operatorname{cov}({{X}_{2}},Y)
cov(X1+X2,Y)=cov(X1,Y)+cov(X2,Y),常数消除剩下四项,再通过
D
X
=
cov
(
X
,
X
)
DX=\operatorname{cov}(X,X)
DX=cov(X,X)和
cov
(
X
,
Y
)
=
cov
(
Y
,
X
)
\operatorname{cov}(X,Y)=\operatorname{cov}(Y,X)
cov(X,Y)=cov(Y,X)化简
相关系数
ρ X Y = cov ( X , Y ) D X ⋅ D Y . {{\rho }_{XY}}\\=\frac{\operatorname{cov}(X,Y)}{\sqrt{DX}\cdot \sqrt{DY}}. ρXY=DX⋅DYcov(X,Y).
= E [ ( X − E X ) ( Y − E Y ) ] D X ⋅ D Y =\frac{E[(X-EX)(Y-EY)]}{\sqrt{DX}\cdot \sqrt{DY}} =DX⋅DYE[(X−EX)(Y−EY)]
= E [ ( X − E X D X ) ( Y − E Y D Y ) ] =E[(\frac{X-EX}{\sqrt{DX}})(\frac{Y-EY}{\sqrt{DY}})] =E[(DXX−EX)(DYY−EY)]
性质
X,Y相互独立,则有
ρ
=
0.
\rho =0.
ρ=0.
若
ρ
X
Y
=
cov
(
X
,
Y
)
=
0
{{\rho }_{XY}}=\operatorname{cov}(X,Y)=0
ρXY=cov(X,Y)=0,则X,Y不相关(不是独立)
独立一定不相关,不相关不一定独立
相互独立的正态分布之和
ξ ~ N ( μ 1 , σ 1 2 ) , η ~ N ( μ 2 , σ 2 2 ) , \xi \tilde{\ }N({{\mu }_{1}},\sigma _{1}^{2}),\eta \tilde{\ }N({{\mu }_{2}},\sigma _{2}^{2}), ξ ~N(μ1,σ12),η ~N(μ2,σ22),且 ξ , η \xi ,\eta ξ,η相互独立 则有 ( ξ + η ) ~ N ( μ 1 + μ 2 , σ 1 2 + σ 2 2 ) (\xi +\eta) \tilde{\ }N({{\mu }_{1}}+{{\mu }_{2}},\sigma _{1}^{2}+\sigma _{2}^{2}) (ξ+η) ~N(μ1+μ2,σ12+σ22)
二维正态分布
(
X
,
Y
)
~
N
(
μ
1
,
μ
2
,
σ
1
2
,
σ
2
2
,
ρ
)
,
(X,Y)\tilde{\ }N({{\mu }_{1}},{{\mu }_{2}},\sigma _{1}^{2},\sigma _{2}^{2},\rho ),
(X,Y) ~N(μ1,μ2,σ12,σ22,ρ),
协方差
cov
(
X
,
Y
)
=
ρ
σ
1
σ
2
\operatorname{cov}(X,Y)=\rho {{\sigma }_{1}}{{\sigma }_{2}}
cov(X,Y)=ρσ1σ2
相关系数
ρ
X
Y
=
C
o
v
(
X
,
Y
)
D
X
⋅
D
Y
=
ρ
.
{{\rho }_{XY}}=\frac{Cov(X,Y)}{\sqrt{DX}\cdot \sqrt{DY}}=\rho .
ρXY=DX⋅DYCov(X,Y)=ρ.
ρ
=
0
⇔
X
Y
不
相
关
且
相
互
独
立
\rho=0\Leftrightarrow XY不相关且相互独立
ρ=0⇔XY不相关且相互独立
卷积公式
用于求解已知f(x,y),求解
f
Z
(
z
)
f_Z(z)
fZ(z)
f
Z
(
z
)
=
∫
−
∞
+
∞
f
(
x
,
z
−
x
)
d
x
{{f}_{Z}}(z)=\int_{-\infty }^{+\infty }{f}(x,z-x)dx
fZ(z)=∫−∞+∞f(x,z−x)dx
切比雪夫不等式
P { ∣ ξ − μ ∣ ≥ ε } ≤ D( ξ ) ε 2 ⇔ P { ∣ ξ − μ ∣ < ε } ≥ 1 − D( ξ ) ε 2 P\{\left| \xi -\mu \right|\ge \varepsilon \}\le \frac{\text{D(}\xi {{\text{)}}^{{}}}}{{{\varepsilon }^{2}}}\Leftrightarrow P\{\left| \xi -\mu \right|<\varepsilon \}\ge 1-\frac{\text{D(}\xi \text{)}}{{{\varepsilon }^{2}}} P{∣ξ−μ∣≥ε}≤ε2D(ξ)⇔P{∣ξ−μ∣<ε}≥1−ε2D(ξ)
当 ε 取 k D ε 时 有 μ = E ( ξ ) \varepsilon 取k\sqrt{D\varepsilon } 时有\mu =E(\xi ) ε取kDε时有μ=E(ξ)
P { ∣ ξ − μ ∣ ≥ k D ξ } ≤ 1 k 2 P\{\left| \xi -\mu \right|\ge k\sqrt{D\xi }\}\le \frac{1}{{{k}^{2}}} P{∣ξ−μ∣≥kDξ}≤k21 or P { ∣ ξ − μ ∣ k D ξ } ≥ 1- 1 k 2 P\{\left| \xi -\mu \right|\text{}k\sqrt{D\xi }\}\ge \text{1-}\frac{1}{{{k}^{2}}} P{∣ξ−μ∣kDξ}≥1-k21
德莫佛-拉普拉斯定理
即设随机变量
X
(
n
=
1
,
2
,
.
.
.
,
)
X(n=1,2,...,)
X(n=1,2,...,)服从参数为
n
,
p
(
0
<
p
<
1
)
n,p(0<p<1)
n,p(0<p<1)的二项分布,则对于任意有限区间
(
a
,
b
)
(a,b)
(a,b)有
lim
n
→
∞
P
{
a
≤
x
n
−
n
p
n
p
(
1
−
p
)
≤
b
}
=
∫
a
b
1
2
π
e
−
t
2
2
d
t
\lim_{n \to \infty} P\{a\le \frac{x_n-np}{\sqrt{np(1-p)}} \le b\}=\int_{a}^{b}\frac{1}{\sqrt{2\pi}}e^-{\frac{t^2}{2}dt}
limn→∞P{a≤np(1−p)xn−np≤b}=∫ab2π1e−2t2dt
正态分布是二项分布的极限分布,所以当n充分大时,我们可以用标准正态分布近似二项分布.
基本函数求导
- ( C ) ′ = 0 (C)'=0 (C)′=0
- ( x a ) ′ = a x a − 1 (x^a)'=ax^{a-1} (xa)′=axa−1
- ( s i n x ) ′ = c o s x (sinx)'=cosx (sinx)′=cosx
- ( c o s x ) ′ = − s i n x (cosx)'=-sinx (cosx)′=−sinx
- ( t a n x ) ′ = s e c 2 x (tanx)'=sec^2x (tanx)′=sec2x
- ( c o t x ) ′ = − c s c 2 x (cotx)'=-csc^2x (cotx)′=−csc2x
- ( s e c x ) ′ = s e c x t a n x (secx)'=secxtanx (secx)′=secxtanx
- ( c s c x ) ′ = − c s c x c o t x (cscx)'=-cscxcotx (cscx)′=−cscxcotx
- ( a x ) ′ = a x l n a (a^x)'=a^xlna (ax)′=axlna
- ( e x ) ′ = e x (e^x)'=e^x (ex)′=ex
- ( l o g a x ) ′ = 1 x l n a (log_ax)'=\frac{1}{xlna} (logax)′=xlna1
- ( l n x ) ′ = 1 x (lnx)'=\frac{1}{x} (lnx)′=x1
- ( a r c s i n x ) ′ = 1 1 − x 2 (arcsinx)'=\frac{1}{\sqrt{1-x^2}} (arcsinx)′=1−x21
- ( a r c c o s x ) ′ = − 1 1 − x 2 (arccosx)'=-\frac{1}{\sqrt{1-x^2}} (arccosx)′=−1−x21
- ( a r c t a n x ) ′ = 1 1 + x 2 (arctanx)'=\frac{1}{1+x^2} (arctanx)′=1+x21
- ( a r c c o t x ) ′ = − 1 1 + x 2 (arccotx)'=-\frac{1}{1+x^2} (arccotx)′=−1+x21