应用多元统计分析(2022春)
3.1
设 x ∼ N p ( μ , Σ ) , rank ( Σ ) = r , u ∼ N r ( 0 , I ) \boldsymbol{x} \sim N_{p}(\boldsymbol{\mu}, \boldsymbol{\Sigma}), \operatorname{rank}(\boldsymbol{\Sigma})=r, \boldsymbol{u} \sim N_{r}(\boldsymbol{0}, \boldsymbol{I}) x∼Np(μ,Σ),rank(Σ)=r,u∼Nr(0,I), 试证存在秩为 r r r 的 p × r p \times r p×r 矩阵 A \boldsymbol{A} A, 使得 x \boldsymbol{x} x 和 μ + A u \boldsymbol{\mu}+\boldsymbol{A} \boldsymbol{u} μ+Au 服从相同的分布。
答案
由 § 1.7 \S 1.7 §1.7 的性质 (8) 知, 存在秩为 r r r 的 p × r p \times r p×r 矩阵 A \boldsymbol{A} A, 使得 Σ = A A ′ \boldsymbol{\Sigma}=\boldsymbol{A} \boldsymbol{A}^{\prime} Σ=AA′ 。由于 u ∼ N r ( 0 , I ) \boldsymbol{u} \sim N_{r}(\mathbf{0}, \boldsymbol{I}) u∼Nr(0,I), 故 μ + A u ∼ \boldsymbol{\mu}+\boldsymbol{A u \sim} μ+Au∼ N p ( μ , A A ′ ) N_{p}\left(\boldsymbol{\mu}, \boldsymbol{A} \boldsymbol{A}^{\prime}\right) Np(μ,AA′), 从而 μ + A u \boldsymbol{\mu}+\boldsymbol{A} \boldsymbol{u} μ+Au 与 x \boldsymbol{x} x 服从相同的分布 N p ( μ , Σ ) N_{p}(\boldsymbol{\mu}, \boldsymbol{\Sigma}) Np(μ,Σ) 。
3.2
设
x
∼
N
3
(
μ
,
Σ
)
\boldsymbol{x} \sim N_{3}(\boldsymbol{\mu}, \boldsymbol{\Sigma})
x∼N3(μ,Σ), 其中
Σ
=
(
4
−
2
−
2
−
2
2
0
−
2
0
2
)
\boldsymbol{\Sigma}=\left(\begin{array}{rrr} 4 & -2 & -2 \\ -2 & 2 & 0 \\ -2 & 0 & 2 \end{array}\right)
Σ=⎝
⎛4−2−2−220−202⎠
⎞
试求习题 3. 1中的
A
\boldsymbol{A}
A 。
答案
Σ
\Sigma
Σ 的特征值分别为
6
,
2
,
0
6,2,0
6,2,0, 前两个正特征值相应的单位特征向量分别为
(
2
/
6
,
−
1
/
6
,
−
1
/
6
)
′
(2 / \sqrt{6},-1 / \sqrt{6},-1 / \sqrt{6})^{\prime}
(2/6,−1/6,−1/6)′ 和
(
0
,
1
/
2
,
−
1
/
2
)
′
(0,1 / \sqrt{2},-1 / \sqrt{2})^{\prime}
(0,1/2,−1/2)′, 由
§
1.7
\S 1.7
§1.7 性质 (8) 的证明过程知,
A
=
(
6
(
2
/
6
−
1
/
6
−
1
/
6
)
,
2
(
0
1
/
2
−
1
/
2
)
)
=
(
2
0
−
1
1
−
1
−
1
)
\boldsymbol{A}=\left( \sqrt{6}\left(\begin{array}{r} 2 / \sqrt{6} \\ -1 / \sqrt{6} \\ -1 / \sqrt{6} \end{array}\right), \sqrt{2}\left(\begin{array}{c} 0 \\ 1 / \sqrt{2} \\ -1 / \sqrt{2} \end{array}\right)\right) =\left(\begin{array}{rr} 2 & 0 \\ -1 & 1 \\ -1 & -1 \end{array}\right)
A=⎝
⎛6⎝
⎛2/6−1/6−1/6⎠
⎞,2⎝
⎛01/2−1/2⎠
⎞⎠
⎞=⎝
⎛2−1−101−1⎠
⎞
text = "
4 -2 -2
-2 2 0
-2 0 2
"
Sigma = data.matrix(read.table(text=text, fill=TRUE))
colnames(Sigma) = NULL
Sigma
Lambda = eigen(Sigma)$values
Lambda
## [1] 6.000000e+00 2.000000e+00 3.552714e-15
T = eigen(Sigma)$vectors
r = qr(Sigma)$rank
r
## [1] 2
Lambda1 = diag(Lambda[1:r])
T1 = T[,1:r]
A = T1 %*% sqrt(Lambda1 )
A
## [,1] [,2]
## [1,] 2 0
## [2,] -1 -1
## [3,] -1 1
3.3
设 x ∼ N 3 ( μ , Σ ) , A = ( 1 2 − 1 1 2 − 1 2 0 − 1 2 ) \boldsymbol{x} \sim N_{3}(\boldsymbol{\mu}, \boldsymbol{\Sigma}), \boldsymbol{A}=\left(\begin{array}{rrr}\frac{1}{2} & -1 & \frac{1}{2} \\ -\frac{1}{2} & 0 & -\frac{1}{2}\end{array}\right) x∼N3(μ,Σ),A=(21−21−1021−21), 其中 μ = ( 1 , 2 , − 1 ) ′ , Σ = ( 2 1 1 1 2 − 1 1 − 1 4 ) \boldsymbol{\mu}=(1,2,-1)^{\prime}, \boldsymbol{\Sigma}=\left(\begin{array}{rrr}2 & 1 & 1 \\ 1 & 2 & -1 \\ 1 & -1 & 4\end{array}\right) μ=(1,2,−1)′,Σ=⎝ ⎛21112−11−14⎠ ⎞, 试求 y = A x \boldsymbol{y}=\boldsymbol{A} \boldsymbol{x} y=Ax 的分布。
答案
计算得
E
(
y
)
=
A
μ
=
(
−
2
0
)
,
V
(
y
)
=
A
Σ
A
′
=
(
4
−
2
−
2
2
)
E(\boldsymbol{y})=\boldsymbol{A} \boldsymbol{\mu}=\left(\begin{array}{r}-2 \\ 0\end{array}\right), \quad V(\boldsymbol{y})=\boldsymbol{A} \boldsymbol{\Sigma} \boldsymbol{A}^{\prime}=\left(\begin{array}{rr}4 & -2 \\ -2 & 2\end{array}\right)
E(y)=Aμ=(−20),V(y)=AΣA′=(4−2−22)
所以
y
\boldsymbol{y}
y 服从
N
2
(
(
−
2
0
)
,
(
4
−
2
−
2
2
)
)
N_{2}\left(\left(\begin{array}{r}-2 \\ 0\end{array}\right),\left(\begin{array}{rr}4 & -2 \\ -2 & 2\end{array}\right)\right)
N2((−20),(4−2−22)) 。
A = matrix(c(1/2,-1/2,-1,0,1/2,-1/2),2,3)
mu = c(1,2,-1)
Sigma = matrix(c(2,1,1,1,2,-1,1,-1,4),3,3)
Ey = A %*% mu
Ey
## [,1]
## [1,] -2
## [2,] 0
Vy = A %*% Sigma %*% t(A)
Vy
## [,1] [,2]
## [1,] 4 -2
## [2,] -2 2
3.4
设
x
∼
N
3
(
μ
,
Σ
)
\boldsymbol{x} \sim N_{3}(\boldsymbol{\mu}, \boldsymbol{\Sigma})
x∼N3(μ,Σ), 其中
μ
=
(
3
1
4
)
,
Σ
=
(
6
1
−
2
1
13
4
−
2
4
4
)
\boldsymbol{\mu}=\left(\begin{array}{l} 3 \\ 1 \\ 4 \end{array}\right), \quad \boldsymbol{\Sigma}=\left(\begin{array}{rrr} 6 & 1 & -2 \\ 1 & 13 & 4 \\ -2 & 4 & 4 \end{array}\right)
μ=⎝
⎛314⎠
⎞,Σ=⎝
⎛61−21134−244⎠
⎞
试求
(1)
y
1
=
x
1
+
x
2
−
2
x
3
y_{1}=x_{1}+x_{2}-2 x_{3}
y1=x1+x2−2x3 和
y
2
=
3
x
1
−
x
2
+
2
x
3
y_{2}=3 x_{1}-x_{2}+2 x_{3}
y2=3x1−x2+2x3 的联合分布;
(2)
x
1
x_{1}
x1 和
x
3
x_{3}
x3 的联合分布;
(3)
x
1
,
x
3
x_{1}, x_{3}
x1,x3 和
1
2
(
x
1
+
x
2
)
\frac{1}{2}\left(x_{1}+x_{2}\right)
21(x1+x2) 的联合分布。
答案
(1)
计算得
( y 1 y 2 ) = ( 1 1 − 2 3 − 1 2 ) ( x 1 x 2 x 3 ) ∼ N 2 ( ( − 4 16 ) , ( 29 15 15 37 ) ) \left(\begin{array}{l}y_{1} \\ y_{2}\end{array}\right) =\left(\begin{array}{rrr} 1 & 1 & -2 \\ 3 & -1 & 2\end{array}\right) \left(\begin{array}{l}x_{1} \\ x_{2} \\ x_{3}\end{array}\right) \sim N_{2}\left(\left(\begin{array}{r}-4 \\ 16\end{array}\right), \quad \left(\begin{array}{ll} 29 & 15 \\ 15 & 37\end{array}\right)\right) (y1y2)=(131−1−22)⎝ ⎛x1x2x3⎠ ⎞∼N2((−416),(29151537))
A = matrix(c(1,3,1,-1,-2,2),2,3)
mu = c(3,1,4)
Sigma = matrix(c(6, 1, -2, 1, 13, 4, -2, 4, 4), nrow=3)
Ey = A %*% mu; Ey
## [,1]
## [1,] -4
## [2,] 16
Vy = A %*% Sigma %*% t(A); Vy
## [,1] [,2]
## [1,] 29 15
## [2,] 15 37
(2)
注意到
(
x
1
x
3
)
=
(
1
0
0
0
0
1
)
(
x
1
x
2
x
3
)
=
def
B
x
\left(\begin{array}{l}x_{1} \\ x_{3}\end{array}\right) =\left(\begin{array}{rrr} 1 & 0 & 0 \\ 0 & 0 & 1\end{array}\right) \left(\begin{array}{l}x_{1} \\ x_{2} \\ x_{3}\end{array}\right) \overset{\text{def}}=B \boldsymbol{x}
(x1x3)=(100001)⎝
⎛x1x2x3⎠
⎞=defBx
仿照(1)计算可得
N
2
(
(
3
4
)
,
(
6
−
2
−
2
4
)
)
;
N_{2}\left(\left(\begin{array}{l}3 \\ 4\end{array}\right),\left(\begin{array}{rr}6 & -2 \\ -2 & 4\end{array}\right)\right) ;
N2((34),(6−2−24));
B = matrix(c(1,0,0,0,0,1),2,3)
mu = c(3,1,4)
Sigma = matrix(c(6, 1, -2, 1, 13, 4, -2, 4, 4), nrow=3)
Ey = B %*% mu; Ey
## [,1]
## [1,] 3
## [2,] 4
Vy = B %*% Sigma %*% t(B); Vy
## [,1] [,2]
## [1,] 6 -2
## [2,] -2 4
(3)
( x 1 x 3 1 2 ( x 1 + x 2 ) ) = ( 1 0 0 0 0 1 1 2 1 2 0 ) ( x 1 x 2 x 3 ) ∼ N 3 ( ( 3 4 2 ) , ( 6 − 2 3 1 2 − 2 4 1 3 1 2 1 5 1 4 ) ) \left(\begin{array}{c} x_{1} \\ x_{3} \\ \frac{1}{2}\left(x_{1}+x_{2}\right)\end{array}\right) =\left(\begin{array}{lll} 1 & 0 & 0 \\ 0 & 0 & 1 \\ \frac{1}{2} & \frac{1}{2} & 0\end{array}\right) \left(\begin{array}{l}x_{1} \\ x_{2} \\ x_{3}\end{array}\right) \sim N_{3} \left(\left(\begin{array}{l} 3 \\ 4 \\ 2\end{array}\right), \left(\begin{array}{ccc} 6 & -2 & 3\frac{1}{2} \\ -2 & 4 & 1 \\ 3\frac{1}{2} & 1 & 5\frac{1}{4} \end{array}\right)\right) ⎝ ⎛x1x321(x1+x2)⎠ ⎞=⎝ ⎛10210021010⎠ ⎞⎝ ⎛x1x2x3⎠ ⎞∼N3⎝ ⎛⎝ ⎛342⎠ ⎞,⎝ ⎛6−2321−2413211541⎠ ⎞⎠ ⎞
C = matrix(c(1,0,1/2,0,0,1/2,0,1,0),3)
mu = c(3,1,4)
Sigma = matrix(c(6, 1, -2, 1, 13, 4, -2, 4, 4), nrow=3)
Ey = C %*% mu; Ey
## [,1]
## [1,] 3
## [2,] 4
## [3,] 2
Vy = C %*% Sigma %*% t(C); Vy
## [,1] [,2] [,3]
## [1,] 6.0 -2 3.50
## [2,] -2.0 4 1.00
## [3,] 3.5 1 5.25
3.5
设 x ∼ N p ( μ , Σ ) \boldsymbol{x} \sim N_{p}(\boldsymbol{\mu}, \boldsymbol{\Sigma}) x∼Np(μ,Σ), 其中 x = ( x 1 , x 2 , ⋯ , x p ) ′ , μ = ( μ 1 , μ 2 , ⋯ , μ p ) ′ , Σ = diag ( σ 1 2 , σ 2 2 , ⋯ , σ p 2 ) \boldsymbol{x}=\left(x_{1}, x_{2}, \cdots, x_{p}\right)^{\prime}, \boldsymbol{\mu}=\left(\mu_{1}, \mu_{2}, \cdots, \mu_{p}\right)^{\prime}, \boldsymbol{\Sigma}=\operatorname{diag}\left(\sigma_{1}^{2}, \sigma_{2}^{2}, \cdots, \sigma_{p}^{2}\right) x=(x1,x2,⋯,xp)′,μ=(μ1,μ2,⋯,μp)′,Σ=diag(σ12,σ22,⋯,σp2), 试证 x 1 x_{1} x1, x 2 , ⋯ , x p x_{2}, \cdots, x_{p} x2,⋯,xp 相互独立。
答案
由
x
∼
N
p
(
μ
,
Σ
)
\boldsymbol{x} \sim N_{p}(\boldsymbol{\mu}, \boldsymbol{\Sigma})
x∼Np(μ,Σ) 可知,
x
i
∼
N
(
μ
i
,
σ
i
2
)
x_{i} \sim N\left(\mu_{i}, \sigma_{i}^{2}\right)
xi∼N(μi,σi2) 。由于
f
(
x
)
=
(
2
π
)
−
p
/
2
∣
Σ
∣
−
1
/
2
exp
[
−
1
2
(
x
−
μ
)
′
Σ
−
1
(
x
−
μ
)
]
=
(
2
π
)
−
p
/
2
(
σ
1
σ
2
⋯
σ
p
)
−
1
exp
[
−
1
2
∑
i
=
1
p
(
x
i
−
μ
i
σ
i
)
2
]
=
∏
i
=
1
p
1
2
π
σ
i
e
(
x
i
−
μ
i
)
2
2
σ
i
2
=
∏
i
=
1
p
f
i
(
x
i
)
\begin{aligned} f(\boldsymbol{x})&=(2 \pi)^{-p / 2}|\boldsymbol{\Sigma}|^{-1 / 2} \exp \left[-\frac{1}{2}(\boldsymbol{x}-\boldsymbol{\mu})^{\prime} \boldsymbol{\Sigma}^{-1}(\boldsymbol{x}-\boldsymbol{\mu})\right] \\ &=(2 \pi)^{-p / 2}\left(\sigma_{1} \sigma_{2} \cdots \sigma_{p}\right)^{-1} \exp \left[-\frac{1}{2} \sum_{i=1}^{p}\left(\frac{x_{i}-\mu_{i}}{\sigma_{i}}\right)^{2}\right] \\ &=\prod_{i=1}^{p} \frac{1}{\sqrt{2 \pi} \sigma_{i}} \mathrm{e}^{\frac{\left(x_{i}-\mu_{i}\right)^{2}}{2 \sigma_{i}^{2}}}=\prod_{i=1}^{p} f_{i}\left(x_{i}\right) \end{aligned}
f(x)=(2π)−p/2∣Σ∣−1/2exp[−21(x−μ)′Σ−1(x−μ)]=(2π)−p/2(σ1σ2⋯σp)−1exp[−21i=1∑p(σixi−μi)2]=i=1∏p2πσi1e2σi2(xi−μi)2=i=1∏pfi(xi)
联合密度等于边缘密度乘积, 故 x 1 , x 2 , ⋯ , x p x_{1}, x_{2}, \cdots, x_{p} x1,x2,⋯,xp 相互独立。
3.6(有用结论)
试证独立正态变量的联合分布必然是多元正态的。
答案
设
x
i
∼
N
(
μ
i
,
σ
i
2
)
x_{i} \sim N\left(\mu_{i}, \sigma_{i}^{2}\right)
xi∼N(μi,σi2), 且
x
1
,
x
2
,
⋯
,
x
p
x_{1}, x_{2}, \cdots, x_{p}
x1,x2,⋯,xp 相互独立, 则
f
(
x
1
,
x
2
,
⋯
,
x
p
)
=
∏
i
=
1
p
f
i
(
x
i
)
=
(
2
π
)
−
p
/
2
(
σ
1
σ
2
⋯
σ
p
)
−
1
exp
[
−
1
2
∑
i
=
1
p
(
x
i
−
μ
i
σ
i
)
2
]
=
(
2
π
)
−
p
/
2
∣
Σ
∣
−
1
/
2
exp
[
−
1
2
(
x
−
μ
)
′
Σ
−
1
(
x
−
μ
)
]
\begin{aligned} f\left(x_{1}, x_{2}, \cdots, x_{p}\right) &=\prod_{i=1}^{p} f_{i}\left(x_{i}\right)=(2 \pi)^{-p / 2}\left(\sigma_{1} \sigma_{2} \cdots \sigma_{p}\right)^{-1} \exp \left[-\frac{1}{2} \sum_{i=1}^{p}\left(\frac{x_{i}-\mu_{i}}{\sigma_{i}}\right)^{2}\right] \\ &=(2 \pi)^{-p / 2}|\boldsymbol{\Sigma}|^{-1 / 2} \exp \left[-\frac{1}{2}(\boldsymbol{x}-\boldsymbol{\mu})^{\prime} \boldsymbol{\Sigma}^{-1}(\boldsymbol{x}-\boldsymbol{\mu})\right] \end{aligned}
f(x1,x2,⋯,xp)=i=1∏pfi(xi)=(2π)−p/2(σ1σ2⋯σp)−1exp[−21i=1∑p(σixi−μi)2]=(2π)−p/2∣Σ∣−1/2exp[−21(x−μ)′Σ−1(x−μ)]
其中
Σ
=
diag
(
σ
1
2
,
σ
2
2
,
⋯
,
σ
p
2
)
>
0
\boldsymbol{\Sigma}=\operatorname{diag}\left(\sigma_{1}^{2}, \sigma_{2}^{2}, \cdots, \sigma_{p}^{2}\right)>0
Σ=diag(σ12,σ22,⋯,σp2)>0, 所以
x
1
,
x
2
,
⋯
,
x
p
x_{1}, x_{2}, \cdots, x_{p}
x1,x2,⋯,xp 的联合分布是多元正态的。
3.7
设
x
∼
N
4
(
μ
,
Σ
)
\boldsymbol{x} \sim N_{4}(\boldsymbol{\mu}, \boldsymbol{\Sigma})
x∼N4(μ,Σ), 其中
μ
=
(
−
4
2
5
−
1
)
,
Σ
=
(
8
0
−
1
0
0
3
0
2
−
1
0
5
0
0
2
0
7
)
\boldsymbol{\mu}=\left(\begin{array}{r} -4 \\ 2 \\ 5 \\ -1\end{array}\right), \quad \boldsymbol{\Sigma}=\left(\begin{array}{rrrr} 8 & 0 & -1 & 0 \\ 0 & 3 & 0 & 2 \\ -1 & 0 & 5 & 0 \\ 0 & 2 & 0 & 7 \end{array}\right)
μ=⎝
⎛−425−1⎠
⎞,Σ=⎝
⎛80−100302−10500207⎠
⎞
以下哪些随机变量对是独立的? 为什么?
(1)
x
1
x_{1}
x1 和
x
2
x_{2}
x2; (2)
x
1
x_{1}
x1 和
x
3
x_{3}
x3; (3)
x
2
x_{2}
x2 和
x
3
x_{3}
x3; (4)
x
3
x_{3}
x3 和
x
4
x_{4}
x4; (5)
(
x
1
,
x
2
)
\left(x_{1}, x_{2}\right)
(x1,x2) 和
x
3
x_{3}
x3; (6)
(
x
1
,
x
3
)
\left(x_{1}, x_{3}\right)
(x1,x3) 和
x
4
;
x_{4} ;
x4; (7)
x
1
x_{1}
x1 和
(
x
2
\left(x_{2}\right.
(x2,
x
4
)
\left.x_{4}\right)
x4); (8)
(
x
1
,
x
2
)
\left(x_{1}, x_{2}\right)
(x1,x2) 和
(
x
3
,
x
4
)
\left(x_{3}, x_{4}\right)
(x3,x4); (9)
(
x
1
,
x
3
)
\left(x_{1}, x_{3}\right)
(x1,x3) 和
(
x
2
,
x
4
)
\left(x_{2}, x_{4}\right)
(x2,x4) 。
答案
注: 答案提示只给出了(1), (3), (4), (6), (7) 和 (9)是独立的, 但没有给出理由, 请补充.
(1)
( x 1 x 2 ) = ( 1 0 0 1 ) ( x 1 x 2 ) C o v ( x 1 , x 2 ) = ( 1 , 0 ) ( 8 0 0 3 ) ( 0 1 ) = 0 \left(\begin{array}{c}x_{1} \\ x_{2}\end{array}\right)=\left(\begin{array}{cc}1&0 \\0&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{2}\end{array}\right)\\ Cov(x_{1},x_{2})=\left(1,0\right)\left(\begin{array}{cc}8&0\\0&3\end{array}\right)\left(\begin{array}{c}0\\1\end{array}\right)=0 (x1x2)=(1001)(x1x2)Cov(x1,x2)=(1,0)(8003)(01)=0
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 x_{1} x1 和 x 2 x_{2} x2 相互独立
(2)
( x 1 x 3 ) = ( 1 0 0 1 ) ( x 1 x 3 ) C o v ( x 1 , x 3 ) = ( 1 , 0 ) ( 8 − 1 − 1 5 ) ( 0 1 ) = − 1 \left(\begin{array}{c}x_{1} \\ x_{3}\end{array}\right)=\left(\begin{array}{cc}1&0 \\0&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{3}\end{array}\right)\\ Cov(x_{1},x_{3})=\left(1,0\right)\left(\begin{array}{cc}8&-1\\-1&5\end{array}\right)\left(\begin{array}{c}0\\1\end{array}\right)=-1 (x1x3)=(1001)(x1x3)Cov(x1,x3)=(1,0)(8−1−15)(01)=−1
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 x_{1} x1 和 x 3 x_{3} x3 不相互独立
(3)
( x 2 x 3 ) = ( 1 0 0 1 ) ( x 2 x 3 ) C o v ( x 2 , x 3 ) = ( 1 , 0 ) ( 3 0 0 5 ) ( 0 1 ) = 0 \left(\begin{array}{c}x_{2} \\ x_{3}\end{array}\right)=\left(\begin{array}{cc}1&0 \\0&1\end{array}\right)\left(\begin{array}{c}x_{2}\\x_{3}\end{array}\right)\\Cov(x_{2},x_{3})=(1,0)\left(\begin{array}{cc}3&0\\0&5\end{array}\right)\left(\begin{array}{c}0\\1\end{array}\right)=0 (x2x3)=(1001)(x2x3)Cov(x2,x3)=(1,0)(3005)(01)=0
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 2 x_{2} x2 和 x 3 x_{3} x3 相互独立
(4)
( x 3 x 4 ) = ( 1 0 0 1 ) ( x 3 x 4 ) C o v ( x 3 , x 4 ) = ( 1 , 0 ) ( 5 0 0 7 ) ( 0 1 ) = 0 \left(\begin{array}{c}x_{3} \\ x_{4}\end{array}\right)=\left(\begin{array}{cc}1&0 \\0&1\end{array}\right)\left(\begin{array}{c}x_{3}\\x_{4}\end{array}\right)\\Cov(x_{3},x_{4})=(1,0)\left(\begin{array}{cc}5&0\\0&7\end{array}\right)\left(\begin{array}{c}0\\1\end{array}\right)=0 (x3x4)=(1001)(x3x4)Cov(x3,x4)=(1,0)(5007)(01)=0
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 3 x_{3} x3 和 x 4 x_{4} x4 相互独立
(5)
( x 1 + x 2 x 3 ) = ( 1 1 0 0 0 1 ) ( x 1 x 2 x 3 ) C o v ( x 1 + x 2 , x 3 ) = ( 1 , 1 , 0 ) ( 8 0 − 1 0 3 0 − 1 0 5 ) ( 0 0 1 ) = − 1 \left(\begin{array}{c}x_{1}+x_{2} \\ x_{3}\end{array}\right)=\left(\begin{array}{ccc}1&1&0 \\ 0&0&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{2}\\x_{3}\end{array}\right)\\Cov(x_{1}+x_{2},x_{3})=(1,1,0)\left(\begin{array}{ccc}8&0&-1\\0&3&0\\-1&0&5\end{array}\right)\left(\begin{array}{c}0\\0\\1\end{array}\right)=-1 (x1+x2x3)=(101001)⎝ ⎛x1x2x3⎠ ⎞Cov(x1+x2,x3)=(1,1,0)⎝ ⎛80−1030−105⎠ ⎞⎝ ⎛001⎠ ⎞=−1
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 + x 2 x_{1}+x_{2} x1+x2 和 x 3 x_{3} x3 相互不独立
(6)
( x 1 + x 3 x 4 ) = ( 1 1 0 0 0 1 ) ( x 1 x 3 x 4 ) C o v ( x 1 + x 3 , x 4 ) = ( 1 , 1 , 0 ) ( 8 − 1 0 − 1 5 0 0 0 7 ) ( 0 0 1 ) = 0 \left(\begin{array}{c}x_{1}+x_{3} \\ x_{4}\end{array}\right)=\left(\begin{array}{ccc}1&1&0 \\ 0&0&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{3}\\x_{4}\end{array}\right)\\Cov(x_{1}+x_{3},x_{4})=(1,1,0)\left(\begin{array}{ccc}8&-1&0\\-1&5&0\\0&0&7\end{array}\right)\left(\begin{array}{c}0\\0\\1\end{array}\right)=0 (x1+x3x4)=(101001)⎝ ⎛x1x3x4⎠ ⎞Cov(x1+x3,x4)=(1,1,0)⎝ ⎛8−10−150007⎠ ⎞⎝ ⎛001⎠ ⎞=0
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 + x 3 x_{1}+x_{3} x1+x3 和 x 4 x_{4} x4 相互独立
(7)
( x 1 x 2 + x 4 ) = ( 1 0 0 0 1 1 ) ( x 1 x 2 x 4 ) C o v ( x 1 , x 2 + x 4 ) = ( 1 , 0 , 0 ) ( 8 0 0 0 3 2 0 2 7 ) ( 0 1 1 ) = 0 \left(\begin{array}{c}x_{1} \\x_{2}+ x_{4}\end{array}\right)=\left(\begin{array}{ccc}1&0&0 \\ 0&1&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{2}\\x_{4}\end{array}\right)\\Cov(x_{1},x_{2}+x_{4})=(1,0,0)\left(\begin{array}{ccc}8&0&0\\0&3&2\\0&2&7\end{array}\right)\left(\begin{array}{c}0\\1\\1\end{array}\right)=0 (x1x2+x4)=(100101)⎝ ⎛x1x2x4⎠ ⎞Cov(x1,x2+x4)=(1,0,0)⎝ ⎛800032027⎠ ⎞⎝ ⎛011⎠ ⎞=0
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 x_{1} x1 和 x 2 + x 4 x_{2}+x_{4} x2+x4 相互独立
(8)
( x 1 + x 2 x 3 + x 4 ) = ( 1 1 0 0 0 0 1 1 ) ( x 1 x 2 x 3 x 4 ) C o v ( x 1 + x 2 , x 3 + x 4 ) = ( 1 , 1 , 0 , 0 ) ( 8 0 − 1 0 0 3 0 2 − 1 0 5 0 0 2 0 7 ) ( 0 0 1 1 ) = 1 \left(\begin{array}{c}x_{1}+x_{2} \\ x_{3}+x_{4}\end{array}\right)=\left(\begin{array}{cccc}1&1&0&0 \\ 0&0&1&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{2}\\x_{3}\\x_{4}\end{array}\right)\\Cov(x_{1}+x_{2},x_{3}+x_{4})=(1,1,0,0)\left(\begin{array}{rrrr}8 & 0 & -1 & 0 \\ 0 & 3 & 0 & 2 \\ -1 & 0 & 5 & 0 \\ 0 & 2 & 0 & 7\end{array}\right)\left(\begin{array}{c}0\\0\\1\\1\end{array}\right)=1 (x1+x2x3+x4)=(10100101)⎝ ⎛x1x2x3x4⎠ ⎞Cov(x1+x2,x3+x4)=(1,1,0,0)⎝ ⎛80−100302−10500207⎠ ⎞⎝ ⎛0011⎠ ⎞=1
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 + x 2 x_{1}+x_{2} x1+x2 和 x 3 + x 4 x_{3}+x_{4} x3+x4 不相互独立
(9)
( x 1 + x 3 x 2 + x 4 ) = ( 1 0 1 0 0 1 0 1 ) ( x 1 x 2 x 3 x 4 ) C o v ( x 1 + x 3 , x 2 + x 4 ) = ( 1 , 0 , 1 , 0 ) ( 8 0 − 1 0 0 3 0 2 − 1 0 5 0 0 2 0 7 ) ( 0 1 0 1 ) = 0 \left(\begin{array}{c}x_{1}+x_{3} \\ x_{2}+x_{4}\end{array}\right)=\left(\begin{array}{cccc}1&0&1&0 \\ 0&1&0&1\end{array}\right)\left(\begin{array}{c}x_{1}\\x_{2}\\x_{3}\\x_{4}\end{array}\right)\\Cov(x_{1}+x_{3},x_{2}+x_{4})=(1,0,1,0)\left(\begin{array}{rrrr}8 & 0 & -1 & 0 \\ 0 & 3 & 0 & 2 \\ -1 & 0 & 5 & 0 \\ 0 & 2 & 0 & 7\end{array}\right)\left(\begin{array}{c}0\\1\\0\\1\end{array}\right)=0 (x1+x3x2+x4)=(10011001)⎝ ⎛x1x2x3x4⎠ ⎞Cov(x1+x3,x2+x4)=(1,0,1,0)⎝ ⎛80−100302−10500207⎠ ⎞⎝ ⎛0101⎠ ⎞=0
从而由 S 3.2 S3.2 S3.2中的性质 ( 6 ) (6) (6)知, x 1 + x 3 x_{1}+x_{3} x1+x3 和 x 2 + x 4 x_{2}+x_{4} x2+x4 相互独立
3.8
设 x ∼ N 2 ( μ , Σ ) \boldsymbol{x} \sim N_{2}(\boldsymbol{\mu}, \boldsymbol{\Sigma}) x∼N2(μ,Σ), 其中 x = ( x 1 , x 2 ) ′ , μ = ( μ 1 , μ 2 ) ′ , Σ = σ 2 ( 1 ρ ρ 1 ) \boldsymbol{x}=\left(x_{1}, x_{2}\right)^{\prime}, \boldsymbol{\mu}=\left(\mu_{1}, \mu_{2}\right)^{\prime}, \boldsymbol{\Sigma}=\sigma^{2}\left(\begin{array}{ll}1 & \rho \\ \rho & 1\end{array}\right) x=(x1,x2)′,μ=(μ1,μ2)′,Σ=σ2(1ρρ1), 试证 x 1 + x 2 x_{1}+x_{2} x1+x2 和 x 1 − x 2 x_{1}-x_{2} x1−x2 相互独立。
答案
因为
x
∼
N
2
(
μ
,
Σ
)
\boldsymbol{x} \sim N_{2}(\boldsymbol{\mu}, \boldsymbol{\Sigma})
x∼N2(μ,Σ),由
y
∼
N
r
(
C
μ
+
b
,
C
Σ
C
′
)
\boldsymbol{y} \sim N_{r}\left(\boldsymbol{C} \boldsymbol{\mu}+\boldsymbol{b}, \boldsymbol{C} \boldsymbol{\Sigma} \boldsymbol{C}^{\prime}\right)
y∼Nr(Cμ+b,CΣC′)得:
(
x
1
+
x
2
x
1
−
x
2
)
=
(
1
1
1
−
1
)
(
x
1
x
2
)
∼
N
2
(
μ
,
Σ
)
\left(\begin{array}{l}x_{1}+x_{2} \\ x_{1}-x_{2}\end{array}\right)=\left(\begin{array}{rr}1 & 1 \\ 1 & -1\end{array}\right)\left(\begin{array}{l}x_{1} \\ x_{2}\end{array}\right) \sim N_{2}(\boldsymbol{\mu}, \boldsymbol{\Sigma})
(x1+x2x1−x2)=(111−1)(x1x2)∼N2(μ,Σ)
又因为
Cov
(
x
1
+
x
2
,
x
1
−
x
2
)
=
Cov
(
(
1
,
1
)
(
x
1
x
2
)
,
(
1
,
−
1
)
(
x
1
x
2
)
)
=
σ
2
(
1
,
1
)
(
1
ρ
ρ
1
)
(
1
−
1
)
=
0
\operatorname{Cov}\left(x_{1}+x_{2}, x_{1}-x_{2}\right)=\operatorname{Cov}\left((1,1)\left(\begin{array}{l}x_{1} \\ x_{2}\end{array}\right),(1,-1)\left(\begin{array}{l}x_{1} \\ x_{2}\end{array}\right)\right)=\sigma^{2}(1,1)\left(\begin{array}{ll}1 & \rho \\ \rho & 1\end{array}\right)\left(\begin{array}{r}1 \\ -1\end{array}\right)=0
Cov(x1+x2,x1−x2)=Cov((1,1)(x1x2),(1,−1)(x1x2))=σ2(1,1)(1ρρ1)(1−1)=0
从而由
S
3.2
S3.2
S3.2中的性质
(
6
)
(6)
(6)知,
x
1
+
x
2
x_{1}+x_{2}
x1+x2 和
x
1
−
x
2
x_{1}-x_{2}
x1−x2 相互独立。