自学内容网 自学内容网

【数学】概率论与数理统计(六)

条件分布


离散型随机变量的条件分布

  • ( X , Y ) (X, Y) (X,Y)是二维离散型随机向量,对于固定的 y j y_{j} yj,若 P {   Y = y j   } > 0 P\set{Y = y_{j}} > 0 P{Y=yj}>0,则称 P {   X = x i ∣ Y = y j   } = P {   X = x i , Y = y j   } P {   Y = y j   } = p i j p ⋅ j , i = 1 , 2 , ⋯ P\set{X = x_{i} | Y = y_{j}} = \frac{P\set{X = x_{i} , Y = y_{j}}}{P\set{Y = y_{j}}} = \frac{p_{ij}}{p_{\cdot j}} , i = 1, 2, \cdots P{X=xiY=yj}=P{Y=yj}P{X=xi,Y=yj}=pjpij,i=1,2,为在 {   Y = y j   } \set{Y = y_{j}} {Y=yj}的条件下,随机变量 X X X的条件分布律,称函数 F ( x ∣ y j ) = P {   X ≤ x ∣ Y = y j   } , x ∈ R F(x | y_{j}) = P\set{X \leq x | Y = y_{j}} , x \in R F(xyj)=P{XxY=yj},xR为在 {   Y = y j   } \set{Y = y_{j}} {Y=yj}的条件下,随机变量 X X X的条件分布函数

示例

问题
  • 一射手进行射击,每次击中目标的概率为 p ( 0 < p < 1 ) p (0 < p < 1) p(0<p<1),且各次射击是相互独立的,射击进行到击中目标两次为止,以 X X X表示首次击中目标所进行的射击次数,以 Y Y Y表示总共进行的射击次数,试求 X X X Y Y Y的联合概率分布律及条件概率分布律
解答
  • {   Y = n   } \set{Y = n} {Y=n}表示在第 n n n次射击时击中目标,且在前 n − 1 n - 1 n1次射击中有一次击中目标
  • 已知各次射击是相互独立的,于是不论 m ( m < n ) m (m < n) m(m<n)是多少,概率 P {   X = m , Y = n   } P\set{X = m , Y = n} P{X=m,Y=n}都应等于 p 2 ( 1 − p ) n − 2 p^{2} (1 - p)^{n - 2} p2(1p)n2,由此得 X X X Y Y Y的联合概率分布律为

P {   X = m , Y = n   } = p 2 ( 1 − p ) n − 2 , n = 2 , 3 , ⋯   , m = 1 , 2 , ⋯   , n − 1 P\set{X = m , Y = n} = p^{2} (1 - p)^{n - 2} , n = 2, 3, \cdots , m = 1, 2, \cdots, n - 1 P{X=m,Y=n}=p2(1p)n2,n=2,3,,m=1,2,,n1

P {   X = m   } = ∑ n = m + 1 ∞ P {   X = m , Y = n   } = ∑ n = m + 1 ∞ p 2 ( 1 − p ) n − 2 = p 2 ∑ n = m + 1 ∞ ( 1 − p ) n − 2 = p 2 ( 1 − p ) m − 1 p = p ( 1 − p ) m − 1 , m = 1 , 2 , ⋯ P\set{X = m} = \sum\limits_{n = m + 1}^{\infty}{P\set{X = m , Y = n}} = \sum\limits_{n = m + 1}^{\infty}{p^{2} (1 - p)^{n - 2}} = p^{2} \sum\limits_{n = m + 1}^{\infty}{(1 - p)^{n - 2}} = \frac{p^{2} (1 - p)^{m - 1}}{p} = p (1 - p)^{m - 1} , m = 1, 2, \cdots P{X=m}=n=m+1P{X=m,Y=n}=n=m+1p2(1p)n2=p2n=m+1(1p)n2=pp2(1p)m1=p(1p)m1,m=1,2,

P {   Y = n   } = ∑ m = 1 n − 1 P {   X = m , Y = n   } = ∑ m = 1 n − 1 p 2 ( 1 − p ) n − 2 = ( n − 1 ) p 2 ( 1 − p ) n − 2 , n = 2 , 3 , ⋯ P\set{Y = n} = \sum\limits_{m = 1}^{n - 1}{P\set{X = m , Y = n}} = \sum\limits_{m = 1}^{n - 1}{p^{2} (1 - p)^{n - 2}} = (n - 1) p^{2} (1 - p)^{n - 2} , n = 2, 3, \cdots P{Y=n}=m=1n1P{X=m,Y=n}=m=1n1p2(1p)n2=(n1)p2(1p)n2,n=2,3,

  • n = 2 , 3 , ⋯ n = 2, 3, \cdots n=2,3,时, P {   X = m ∣ Y = n   } = p 2 ( 1 − p ) n − 2 ( n − 1 ) p 2 ( 1 − p ) n − 2 = 1 n − 1 , m = 1 , 2 , ⋯   , n − 1 P\set{X = m | Y = n} = \frac{p^{2} (1 - p)^{n - 2}}{(n - 1) p^{2} (1 - p)^{n - 2}} = \frac{1}{n - 1} , m = 1, 2, \cdots, n - 1 P{X=mY=n}=(n1)p2(1p)n2p2(1p)n2=n11,m=1,2,,n1
  • m = 1 , 2 , ⋯ m = 1, 2, \cdots m=1,2,时, P {   Y = n ∣ X = m   } = p 2 ( 1 − p ) n − 2 p ( 1 − p ) m − 1 = p ( 1 − p ) n − m − 1 , n = m + 1 , m + 2 , ⋯ P\set{Y = n | X = m} = \frac{p^{2} (1 - p)^{n - 2}}{p (1 - p)^{m - 1}} = p (1 - p)^{n - m - 1} , n = m + 1, m + 2, \cdots P{Y=nX=m}=p(1p)m1p2(1p)n2=p(1p)nm1,n=m+1,m+2,

连续型随机变量的条件分布

  • ( X , Y ) (X, Y) (X,Y)为二维连续型随机向量, y y y取定值,对于任意给定的实数 Δ y > 0 , P {   y < Y ≤ y + Δ y   } > 0 \Delta{y} > 0 , P\set{y < Y \leq y + \Delta{y}} > 0 Δy>0,P{y<Yy+Δy}>0,若极限 lim ⁡ Δ y → 0 + P {   X ≤ x ∣ y < Y ≤ y + Δ y   } \lim\limits_{\Delta{y} \rightarrow 0^{+}}{P\set{X \leq x | y < Y \leq y + \Delta {y}}} Δy0+limP{Xxy<Yy+Δy}存在,则称此极限值为在 {   Y = y   } \set{Y = y} {Y=y}的条件下, X X X的条件分布函数,记为 F ( x ∣ y ) F(x | y) F(xy)
  • 如果二维连续型随机向量 ( X , Y ) (X, Y) (X,Y)的分布函数为 F ( x , y ) F(x, y) F(x,y),概率密度为 f ( x , y ) f(x, y) f(x,y),且 f ( x , y ) f(x, y) f(x,y)在点 ( x , y ) (x, y) (x,y)处连续,而 Y Y Y的边缘概率密度 f Y ( y ) > 0 f_{Y}(y) > 0 fY(y)>0且连续,则

F ( x ∣ y ) = lim ⁡ Δ y → 0 + P {   X ≤ x ∣ y < Y ≤ y + Δ y   } = lim ⁡ Δ y → 0 + P {   X ≤ x , y < Y ≤ y + Δ y   } P {   y < Y ≤ y + Δ y   } = lim ⁡ Δ y → 0 + F ( x , y + Δ y ) − F ( x , y ) F Y ( y + Δ y ) − F Y ( y ) = lim ⁡ Δ y → 0 + F ( x , y + Δ y ) − F ( x , y ) Δ y lim ⁡ Δ y → 0 + F Y ( y + Δ y ) − F Y ( y ) Δ y = ∂ F ( x , y ) ∂ y d d y F Y ( y ) = ∫ − ∞ x f ( u , y ) d u f Y ( y ) \begin{aligned} F(x | y) &= \lim\limits_{\Delta{y} \rightarrow 0^{+}}{P\set{X \leq x | y < Y \leq y + \Delta{y}}} \\ &= \lim\limits_{\Delta{y} \rightarrow 0^{+}}{\cfrac{P\set{X \leq x , y < Y \leq y + \Delta{y}}}{P\set{y < Y \leq y + \Delta{y}}}} \\ &= \lim\limits_{\Delta{y} \rightarrow 0^{+}}{\cfrac{F(x, y + \Delta{y}) - F(x, y)}{F_{Y}(y + \Delta{y}) - F_{Y}(y)}} \\ &= \cfrac{\lim\limits_{\Delta{y} \rightarrow 0^{+}}{\cfrac{F(x, y + \Delta{y}) - F(x, y)}{\Delta{y}}}}{\lim\limits_{\Delta{y} \rightarrow 0^{+}}{\cfrac{F_{Y}(y + \Delta{y}) - F_{Y}(y)}{\Delta{y}}}} \\ &= \cfrac{\cfrac{\partial{F(x, y)}}{\partial{y}}}{\cfrac{d}{dy} F_{Y}(y)} \\ &= \cfrac{\int_{- \infty}^{x}{f(u, y) du}}{f_{Y}(y)} \end{aligned} F(xy)=Δy0+limP{Xxy<Yy+Δy}=Δy0+limP{y<Yy+Δy}P{Xx,y<Yy+Δy}=Δy0+limFY(y+Δy)FY(y)F(x,y+Δy)F(x,y)=Δy0+limΔyFY(y+Δy)FY(y)Δy0+limΔyF(x,y+Δy)F(x,y)=dydFY(y)yF(x,y)=fY(y)xf(u,y)du

  • 即在 {   Y = y   } \set{Y = y} {Y=y}的条件下, X X X的条件分布函数为 F ( x ∣ y ) = ∫ − ∞ x f ( u , y ) d u f Y ( y ) = ∫ − ∞ x f ( u , y ) f Y ( y ) d u F(x | y) = \frac{\int_{- \infty}^{x}{f(u, y) du}}{f_{Y}(y)} = \int_{- \infty}^{x}{\frac{f(u, y)}{f_{Y}(y)} du} F(xy)=fY(y)xf(u,y)du=xfY(y)f(u,y)du
  • 若令 f ( x ∣ y ) = f ( x , y ) f Y ( y ) f(x | y) = \frac{f(x, y)}{f_{Y}(y)} f(xy)=fY(y)f(x,y),则 F ( x ∣ y ) = ∫ − ∞ x f ( u , y ) f Y ( y ) d u = ∫ − ∞ x f ( u ∣ y ) d u F(x | y) = \int_{- \infty}^{x}{\frac{f(u, y)}{f_{Y}(y)} du} = \int_{- \infty}^{x}{f(u | y) du} F(xy)=xfY(y)f(u,y)du=xf(uy)du,称 f ( x ∣ y ) f(x | y) f(xy)为在 {   Y = y   } \set{Y = y} {Y=y}的条件下, X X X的条件概率密度

示例1

问题
  • ( X , Y ) (X, Y) (X,Y)服从区域 D = {   ( x , y ) ∣ x 2 + y 2 ≤ R 2 , R > 0   } D = \set{(x, y) | x^{2} + y^{2} \leq R^{2} , R > 0} D={(x,y)x2+y2R2,R>0}上的均匀分布,求 f ( y ∣ x ) f(y | x) f(yx) f ( x ∣ y ) f(x | y) f(xy)
解答

f ( x , y ) = { 1 π R 2 , ( x , y ) ∈ D 0 , ( x , y ) ∉ D f(x, y) = \begin{cases} \cfrac{1}{\pi R^{2}} , & (x, y) \in D \\ 0 , & (x, y) \notin D \end{cases} f(x,y)= πR21,0,(x,y)D(x,y)/D

f X ( x ) = ∫ − ∞ + ∞ f ( x , y ) d y = { 2 π R 2 R 2 − x 2 , ∣ x ∣ ≤ R 0 , ∣ x ∣ > R f_{X}(x) = \int_{- \infty}^{+ \infty}{f(x, y) dy} = \begin{cases} \cfrac{2}{\pi R^{2}} \sqrt{R^{2} - x^{2}} , & |x| \leq R \\ 0 , & |x| > R \end{cases} fX(x)=+f(x,y)dy= πR22R2x2 ,0,xRx>R

  • 故当 y ∈ ( − R , R ) y \in (- R, R) y(R,R)时,有

f ( x ∣ y ) = f ( x , y ) f Y ( y ) = { 1 2 R 2 − x 2 , ∣ x ∣ ≤ R 2 − y 2 0 , 其他 f(x | y) = \cfrac{f(x, y)}{f_{Y}(y)} = \begin{cases} \cfrac{1}{2 \sqrt{R^{2} - x^{2}}} , & |x| \leq \sqrt{R^{2} - y^{2}} \\ 0 , & 其他 \end{cases} f(xy)=fY(y)f(x,y)= 2R2x2 1,0,xR2y2 其他

  • 可以看出,在 {   Y = y   } \set{Y = y} {Y=y}的条件下, X X X在区间 [ − R 2 − y 2 , R 2 − y 2 ] [- \sqrt{R^{2} - y^{2}}, \sqrt{R^{2} - y^{2}}] [R2y2 ,R2y2 ]上服从均匀分布

示例2

问题
  • 设随机变量 X X X的分布密度为

f X ( x ) = { λ 2 x e − λ x , x > 0 0 , x ≤ 0 f_{X}(x) = \begin{cases} \lambda^{2} x e^{- \lambda x} , & x > 0 \\ 0 , & x \leq 0 \end{cases} fX(x)={λ2xeλx,0,x>0x0

  • 而随机变量 Y Y Y在区间 ( 0 , X ) (0, X) (0,X)上服从均匀分布,求 Y Y Y的概率密度 f Y ( y ) f_{Y}(y) fY(y)
解答

f Y ( y ∣ X = x ) = { 1 x , y ∈ ( 0 , x ) 0 , y ∉ ( 0 , x ) f_{Y}(y | X = x) = \begin{cases} \cfrac{1}{x} , & y \in (0, x) \\ 0 , & y \notin (0, x) \end{cases} fY(yX=x)= x1,0,y(0,x)y/(0,x)

f ( x , y ) = f X ( x ) f Y ( y ∣ X = x ) = { λ 2 e − λ x , 0 < y < x 0 , 其他 f(x, y) = f_{X}(x) f_{Y}(y | X = x) = \begin{cases} \lambda^{2} e^{- \lambda x} , & 0 < y < x \\ 0 , & 其他 \end{cases} f(x,y)=fX(x)fY(yX=x)={λ2eλx,0,0<y<x其他

f Y ( y ) = ∫ − ∞ + ∞ f ( x , y ) d x = { ∫ y + ∞ λ 2 e − λ x d x , y > 0 0 , y ≤ 0 = { λ e − λ y , y > 0 0 , y ≤ 0 f_{Y}(y) = \int_{- \infty}^{+ \infty}{f(x, y) dx} = \begin{cases} \int_{y}^{+ \infty}{\lambda^{2} e^{- \lambda x}} dx , & y > 0 \\ 0 , & y \leq 0 \end{cases} = \begin{cases} \lambda e^{- \lambda y} , & y > 0 \\ 0 , & y \leq 0 \end{cases} fY(y)=+f(x,y)dx={y+λ2eλxdx,0,y>0y0={λeλy,0,y>0y0

  • 即随机变量 Y Y Y服从参数为 λ \lambda λ的指数分布

随机变量的独立性


随机变量相互独立

  • 设二维随机向量 ( X , Y ) (X, Y) (X,Y)的分布函数和边缘分布函数分别为 F ( x , y ) F(x, y) F(x,y) F X ( x ) F_{X}(x) FX(x) F Y ( y ) F_{Y}(y) FY(y),若对于任意的实数 x x x y y y P {   X ≤ x , Y ≤ y   } = P {   X ≤ x   } ⋅ P {   Y ≤ y   } P\set{X \leq x , Y \leq y} = P\set{X \leq x} \cdot P\set{Y \leq y} P{Xx,Yy}=P{Xx}P{Yy} F ( x , y ) = F X ( x ) ⋅ F Y ( y ) F(x, y) = F_{X}(x) \cdot F_{Y}(y) F(x,y)=FX(x)FY(y),则称随机变量 X X X Y Y Y相互独立

  • 设二维离散型随机向量 ( X , Y ) (X, Y) (X,Y)的分布律为 P {   X = x i , Y = y j   } = p i j ( i , j = 1 , 2 , ⋯   ) P\set{X = x_{i} , Y = y_{j}} = p_{ij} (i, j = 1, 2, \cdots) P{X=xi,Y=yj}=pij(i,j=1,2,),关于 X X X Y Y Y的边缘分布律分别为 P {   X = x i   } = p i ⋅ P\set{X = x_{i}} = p_{i \cdot} P{X=xi}=pi P {   Y = y j   } = p ⋅ j P\set{Y = y_{j}} = p_{\cdot j} P{Y=yj}=pj,则 X X X Y Y Y相互独立的充分必要条件是:对任意的 i , j = 1 , 2 , ⋯ i, j = 1, 2, \cdots i,j=1,2,,有 P {   X = x i , Y = y j   } = P {   X = x i   } ⋅ P {   Y = y j   } P\set{X = x_{i} , Y = y_{j}} = P\set{X = x_{i}} \cdot P\set{Y = y_{j}} P{X=xi,Y=yj}=P{X=xi}P{Y=yj},即 p i j = p i ⋅ ⋅ p ⋅ j p_{ij} = p_{i \cdot} \cdot p_{\cdot j} pij=pipj

  • 设二维连续型随机向量 ( X , Y ) (X, Y) (X,Y)的概率密度及边缘概率密度分别为 f ( x , y ) f(x, y) f(x,y) f X ( x ) f_{X}(x) fX(x) f Y ( y ) f_{Y}(y) fY(y),则 X X X Y Y Y相互独立的充要条件是:对于 f ( x , y ) f(x, y) f(x,y) f X ( x ) f_{X}(x) fX(x) f Y ( y ) f_{Y}(y) fY(y)均连续的点 ( x , y ) (x, y) (x,y),有 f ( x , y ) = f X ( x ) ⋅ f Y ( y ) f(x, y) = f_{X}(x) \cdot f_{Y}(y) f(x,y)=fX(x)fY(y)

示例

问题
  • 设二维随机向量 ( X , Y ) ∼ N ( μ 1 , μ 2 , σ 1 2 , σ 2 2 , ρ ) (X, Y) \sim N(\mu_{1}, \mu_{2}, \sigma_{1}^{2}, \sigma_{2}^{2}, \rho) (X,Y)N(μ1,μ2,σ12,σ22,ρ),试证明: X X X Y Y Y相互独立的充分必要条件是 ρ = 0 \rho = 0 ρ=0
解答
  • f ( x , y ) = 1 2 π σ 1 σ 2 1 − ρ 2 exp ⁡ { − 1 2 ( 1 − ρ 2 ) [ ( x − μ 1 ) 2 σ 1 2 − 2 ρ ( x − μ 1 ) ( y − μ 2 ) σ 1 σ 2 + ( y − μ 2 ) 2 σ 2 2 ] } ( − ∞ < x < + ∞ , − ∞ < y < + ∞ ) f(x, y) = \frac{1}{2 \pi \sigma_{1} \sigma_{2} \sqrt{1 - \rho^{2}}} \exp\left\{- \frac{1}{2 (1 - \rho^{2})} \left[\frac{(x - \mu_{1})^{2}}{\sigma_{1}^{2}} - 2 \rho \frac{(x - \mu_{1}) (y - \mu_{2})}{\sigma_{1} \sigma_{2}} + \frac{(y - \mu_{2})^{2}}{\sigma_{2}^{2}}\right]\right\} (- \infty < x < + \infty , - \infty < y < + \infty) f(x,y)=2πσ1σ21ρ2 1exp{2(1ρ2)1[σ12(xμ1)22ρσ1σ2(xμ1)(yμ2)+σ22(yμ2)2]}(<x<+,<y<+)

  • f X ( x ) = 1 2 π σ 1 e − ( x − μ 1 ) 2 2 σ 1 2 , − ∞ < x < + ∞ f_{X}(x) = \frac{1}{\sqrt{2 \pi} \sigma_{1}} e^{- \frac{(x - \mu_{1})^{2}}{2 \sigma_{1}^{2}}} , - \infty < x < + \infty fX(x)=2π σ11e2σ12(xμ1)2,<x<+

  • f Y ( y ) = 1 2 π σ 2 e − ( y − μ 2 ) 2 2 σ 2 2 , − ∞ < y < + ∞ f_{Y}(y) = \frac{1}{\sqrt{2 \pi} \sigma_{2}} e^{- \frac{(y - \mu_{2})^{2}}{2 \sigma_{2}^{2}}} , - \infty < y < + \infty fY(y)=2π σ21e2σ22(yμ2)2,<y<+

  • 先证必要性:若 X X X Y Y Y相互独立,则对于任意的实数 x x x y y y,有 f ( x , y ) = f X ( x ) ⋅ f Y ( y ) f(x, y) = f_{X}(x) \cdot f_{Y}(y) f(x,y)=fX(x)fY(y)

  • x = μ 1 x = \mu_{1} x=μ1 y = μ 2 y = \mu_{2} y=μ2,则有 1 1 − ρ 2 = 1 \frac{1}{\sqrt{1 - \rho^{2}}} = 1 1ρ2 1=1,从而 ρ = 0 \rho = 0 ρ=0

  • 再证充分性:若 ρ = 0 \rho = 0 ρ=0,则对于任意的实数 x x x y y y,必有 f ( x , y ) = f X ( x ) ⋅ f Y ( y ) f(x, y) = f_{X}(x) \cdot f_{Y}(y) f(x,y)=fX(x)fY(y)

  • 因此随机变量 X X X Y Y Y相互独立


定理1

  • 设随机变量 X X X Y Y Y相互独立,又 g ( x ) g(x) g(x) h ( y ) h(y) h(y)是连续函数,则随机变量 g ( X ) g(X) g(X) h ( Y ) h(Y) h(Y)相互独立

定理2

  • ( X 1 , X 2 , ⋯   , X m ) (X_{1}, X_{2}, \cdots, X_{m}) (X1,X2,,Xm) ( Y 1 , Y 2 , ⋯   , Y n ) (Y_{1}, Y_{2}, \cdots, Y_{n}) (Y1,Y2,,Yn)相互独立,则 X i ( i = 1 , 2 , ⋯   , m ) X_{i} (i = 1, 2, \cdots, m) Xi(i=1,2,,m) Y j ( j = 1 , 2 , ⋯   , n ) Y_{j} (j = 1, 2, \cdots, n) Yj(j=1,2,,n)相互独立,又若 g g g h h h是连续函数,则 g ( X 1 , X 2 , ⋯   , X m ) g(X_{1}, X_{2}, \cdots, X_{m}) g(X1,X2,,Xm) h ( Y 1 , Y 2 , ⋯   , Y n ) h(Y_{1}, Y_{2}, \cdots, Y_{n}) h(Y1,Y2,,Yn)相互独立


原文地址:https://blog.csdn.net/from__2025_01_01/article/details/145076612

免责声明:本站文章内容转载自网络资源,如本站内容侵犯了原著者的合法权益,可联系本站删除。更多内容请关注自学内容网(zxcms.com)!