文章目录
- 二、统计量的分布
- 三、正态总体的抽样分布
- 3.1 定理一: X ˉ − μ σ / n ∼ N ( 0 , 1 ) \frac{\bar{X}-\mu}{\sigma/\sqrt{n}}\sim N(0, 1) σ/nXˉ−μ∼N(0,1)( σ \sigma σ已知)
- 3.2 定理二: ( n − 1 ) S 2 σ 2 ∼ χ 2 ( n − 1 ) \frac{(n-1)S^2}{\sigma^2}\sim\chi^2(n-1) σ2(n−1)S2∼χ2(n−1)
- 3.3 定理三: X ˉ − μ S / n ∼ t ( n ) \frac{\bar{X}-\mu}{S/\sqrt{n}}\sim t(n) S/nXˉ−μ∼t(n)( σ \sigma σ未知)
- 3.4 定理四: S 1 2 / S 2 2 σ 1 2 / σ 2 2 ∼ F ( n 1 , n 2 ) \frac{S_1^2/S_2^2}{\sigma_1^2/\sigma_2^2}\sim F(n_1, n_2) σ12/σ22S12/S22∼F(n1,n2)
- 3.5 定理汇总
二、统计量的分布
2.1 统计的基本原理
小概率事件原理:
若事件
A
A
A 发生的概率
P
(
A
)
≤
α
P(A) \leq \alpha
P(A)≤α,则统计规定
A
A
A 在一次抽样下不发生。我们只抽样一次,得
x
ˉ
n
\bar{x}_n
xˉn 用来估计
μ
\mu
μ,于是
μ
\mu
μ 的估计区间应该为:
(
T
1
,
T
2
)
(T_1, T_2)
(T1,T2)(小概率事件抽不到)。即
μ
\mu
μ 的估计区间应该为
X
ˉ
n
∼
N
(
μ
,
σ
2
n
)
\bar{X}_n \sim N(\mu, \frac{\sigma^2}{n})
Xˉn∼N(μ,nσ2) 的非小概率区间端点值
(
T
1
,
T
2
)
(T_1, T_2)
(T1,T2)
P
(
T
1
<
X
ˉ
n
<
T
2
)
=
1
−
α
⟹
X
ˉ
n
−
μ
σ
/
n
∼
N
(
0
,
1
)
P
(
z
1
<
X
ˉ
n
−
μ
σ
/
n
<
z
2
)
=
1
−
α
P
(
X
ˉ
n
−
z
2
σ
n
<
μ
<
X
ˉ
n
−
z
1
σ
n
)
=
1
−
α
P(T_1 < \bar{X}_n < T_2) = 1 - \alpha \quad \Longrightarrow \quad \frac{\bar{X}_n - \mu}{\sigma / \sqrt{n}} \sim N(0, 1) \\ P(z_1 < \frac{\bar{X}_n - \mu}{\sigma / n} < z_2) = 1 - \alpha \\ P(\bar{X}_n - z_2 \frac{\sigma}{\sqrt{n}} < \mu < \bar{X}_n - z_1 \frac{\sigma}{\sqrt{n}}) = 1-\alpha
P(T1<Xˉn<T2)=1−α⟹σ/nXˉn−μ∼N(0,1)P(z1<σ/nXˉn−μ<z2)=1−αP(Xˉn−z2nσ<μ<Xˉn−z1nσ)=1−α
分位点:
(小概率事件区间与非小概率事件区间的分界点
T
1
,
T
2
T_1, T_2
T1,T2)
P
(
X
>
X
α
)
=
α
P(X > X_\alpha) = \alpha
P(X>Xα)=α,
X
α
X_\alpha
Xα 称为随机变量
X
X
X 的上
α
\alpha
α 分位点。用右尾面积的大小标注分位点的位置所在。
2.2 标准正态分布 N ( 0 , 1 ) N(0, 1) N(0,1)
- 定义:
φ ( x ) = 1 2 π e − x 2 2 , − ∞ < x < + ∞ \varphi(x) = \frac{1}{\sqrt{2 \pi}}e^{-\frac{x^2}{2}}, \qquad -\infty<x<+\infty φ(x)=2π1e−2x2,−∞<x<+∞ - 性质: 关于y轴对称。
- 分位点: (上
α
\alpha
α 分位点) (以
α
=
0.05
\alpha=0.05
α=0.05 为例)
- 双侧分位点:
Z
0.975
=
−
Z
0.025
Z_{0.975}=-Z_{0.025}
Z0.975=−Z0.025 及
Z
0.025
Z_{0.025}
Z0.025
- 非小概率事件区间: ( − Z 0.025 , Z 0.025 ) (-Z_{0.025}, Z_{0.025}) (−Z0.025,Z0.025)
- 小概率事件区间: ( − ∞ , − Z 0.025 ) ∪ ( Z 0.025 , ∞ ) (-\infty, -Z_{0.025}) \cup (Z_{0.025}, \infty) (−∞,−Z0.025)∪(Z0.025,∞)
- 单侧上限分位点 (过高异常):
Z
0.05
Z_{0.05}
Z0.05
- 非小概率事件区间: ( − ∞ , Z 0.05 ) (-\infty, Z_{0.05}) (−∞,Z0.05)
- 小概率事件区间: ( Z 0.05 , ∞ ) (Z_{0.05}, \infty) (Z0.05,∞)
- 单侧下限分位点 (过低异常):
Z
0.95
=
−
Z
0.05
Z_{0.95} = -Z_{0.05}
Z0.95=−Z0.05
- 非小概率事件区间: ( − Z 0.05 , ∞ ) (-Z_{0.05}, \infty) (−Z0.05,∞)
- 小概率事件区间: ( − ∞ , − Z 0.05 ) (-\infty, -Z_{0.05}) (−∞,−Z0.05)
- 双侧分位点:
Z
0.975
=
−
Z
0.025
Z_{0.975}=-Z_{0.025}
Z0.975=−Z0.025 及
Z
0.025
Z_{0.025}
Z0.025
例1:
(1)当 α = 0.05 \alpha = 0.05 α=0.05,求 X ∼ N ( 0 , 1 ) X \sim N(0, 1) X∼N(0,1) 双侧分位点。
(2)当 α = 0.01 \alpha = 0.01 α=0.01,求 X ∼ N ( 0 , 1 ) X \sim N(0, 1) X∼N(0,1) 单侧上限分位点。
(3)当 α = 0.1 \alpha = 0.1 α=0.1,求 X ∼ N ( 0 , 1 ) X \sim N(0, 1) X∼N(0,1) 单侧下限分位点。
解:
(1) α = 0.05 \alpha = 0.05 α=0.05 双侧分位点: ± Z 0.025 = ± 1.96 \pm Z_{0.025} = \pm 1.96 ±Z0.025=±1.96
(2) α = 0.01 \alpha = 0.01 α=0.01 单侧上限分位点: Z 0.01 = 2.33 Z_{0.01} = 2.33 Z0.01=2.33
(3) α = 0.1 \alpha = 0.1 α=0.1 单侧下限分位点: Z 0.9 = − Z 0.1 = − 1.28 Z_{0.9} = -Z_{0.1} = -1.28 Z0.9=−Z0.1=−1.28
一个相关的结论(只需了解)
- X ∼ N ( 0 , 1 ) , Y = X 2 X \sim N(0, 1), \; Y = X^2 X∼N(0,1),Y=X2,称 X ∼ χ 2 ( 1 ) X \sim \chi^2(1) X∼χ2(1) 分布,密度函数为:
f Y ( y ) = { 1 2 π y − 1 2 e − y 2 , y > 0 0 , y ≤ 0 = { 1 2 1 2 Γ ( 1 2 ) y 1 2 − 1 e − y 2 , y > 0 0 , y ≤ 0 f_Y(y) = \begin{cases}\frac{1}{\sqrt{2 \pi}}y^{-\frac 12}e^{-\frac y2}, \; \;y > 0 \\ 0, \qquad \qquad \quad y \leq 0 \end{cases} = \begin{cases}\frac 1{2^{\frac12}\Gamma(\frac 12)}y^{\frac 12 -1}e^{- \frac y2}, \; \; y > 0 \\ 0, \qquad \qquad \qquad \; \; y \le 0\end{cases} fY(y)={2π1y−21e−2y,y>00,y≤0=⎩ ⎨ ⎧221Γ(21)1y21−1e−2y,y>00,y≤0- X 1 , X 2 , ⋯ , X n X_1, X_2, \cdots, X_n X1,X2,⋯,Xn 相互独立,且所有 X i ∼ N ( 0 , 1 ) X_i \sim N(0, 1) Xi∼N(0,1)
则 Z = X 1 2 + X 2 2 + ⋯ + X n 2 ∼ χ 2 ( n ) Z = X_1^2 + X_2^2 + \cdots + X_n^2 \sim \chi^2(n) Z=X12+X22+⋯+Xn2∼χ2(n)
f Z ( z ) = { 1 2 n 2 Γ ( n 2 ) z n 2 − 1 e − z 2 , z > 0 0 , z ≤ 0 f_Z(z) = \begin{cases}\frac 1{2^{\frac n2}\Gamma(\frac n2)}z^{\frac n2 -1}e^{- \frac z2}, \; \; z > 0 \\ 0, \qquad \qquad \qquad \; \ \; z \le 0\end{cases} fZ(z)={22nΓ(2n)1z2n−1e−2z,z>00, z≤0 χ 2 ( n ) \chi^2(n) χ2(n) 具有可加性。
2.3 χ 2 ( n ) \chi^2(n) χ2(n) 分布
- 定义:
- X ∼ N ( 0 , 1 ) X \sim N(0, 1) X∼N(0,1) 则称 X 2 ∼ χ 2 ( 1 ) X^2 \sim \chi^2 (1) X2∼χ2(1) 自由度为1的卡方分布。
-
X
1
,
⋯
,
X
n
X_1, \cdots, X_n
X1,⋯,Xn 相互独立,
X
i
∼
N
(
0
,
1
)
,
i
=
1
,
2
,
⋯
,
n
X_i \sim N(0, 1), \quad i=1, 2, \cdots, n
Xi∼N(0,1),i=1,2,⋯,n,则:
∑ i = 1 n X i 2 = χ 2 ( n ) \sum_{i = 1}^nX_i^2 = \chi^2(n) i=1∑nXi2=χ2(n)
- 性质:
- χ 2 ( n ) \chi^2(n) χ2(n) 与 χ 2 ( m ) \chi^2(m) χ2(m) 相互独立,则 χ 2 ( n ) + χ 2 ( m ) ∼ χ 2 ( n + m ) \chi^2(n)+\chi^2(m) \sim \chi^2(n+m) χ2(n)+χ2(m)∼χ2(n+m)
- 若 X ∼ χ 2 ( n ) X \sim \chi^2(n) X∼χ2(n),则 E { χ 2 ( n ) } = n , D { χ 2 ( n ) } = 2 n E\{ \chi^2(n) \} = n, \quad D\{ \chi^2(n) \}=2n E{χ2(n)}=n,D{χ2(n)}=2n
- 分位点: (以
α
=
0.05
\alpha=0.05
α=0.05 为例)
- 双侧分位点:
χ
0.975
2
(
n
)
,
χ
0.025
2
(
n
)
\chi^2_{0.975}(n), \quad \chi^2_{0.025}(n)
χ0.9752(n),χ0.0252(n)
- 非小概率事件区间: { χ 0.975 2 ( n ) , χ 0.025 2 ( n ) } \{ \chi^2_{0.975}(n), \chi^2_{0.025}(n) \} {χ0.9752(n),χ0.0252(n)}
- 小概率事件区间: { 0 , χ 0.975 2 ( n ) } ∪ { χ 0.025 2 ( n ) , + ∞ } \{ 0, \chi^2_{0.975}(n) \} \cup \{ \chi^2_{0.025}(n), +\infty \} {0,χ0.9752(n)}∪{χ0.0252(n),+∞}
- 单侧上限分位点 (过高异常):
χ
0.05
2
(
n
)
\chi^2_{0.05}(n)
χ0.052(n)
- 非小概率事件区间: { 0 , χ 0.05 2 ( n ) } \{ 0, \chi^2_{0.05}(n) \} {0,χ0.052(n)}
- 小概率事件区间: { χ 0.05 2 ( n ) , + ∞ } \{ \chi^2_{0.05}(n), +\infty \} {χ0.052(n),+∞}
- 单侧下限分位点 (过低异常):
χ
0.95
2
(
n
)
\chi^2_{0.95}(n)
χ0.952(n)
- 非小概率事件区间: { χ 0.95 2 ( n ) , + ∞ } \{ \chi^2_{0.95}(n), +\infty \} {χ0.952(n),+∞}
- 小概率事件区间: { 0 , χ 0.95 2 ( n ) } \{ 0, \chi^2_{0.95}(n) \} {0,χ0.952(n)}
- 双侧分位点:
χ
0.975
2
(
n
)
,
χ
0.025
2
(
n
)
\chi^2_{0.975}(n), \quad \chi^2_{0.025}(n)
χ0.9752(n),χ0.0252(n)
例2:
(1)当 α = 0.01 \alpha=0.01 α=0.01,求 X ∼ χ 2 ( 15 ) X \sim \chi^2(15) X∼χ2(15) 双侧分位点。
(2)当 α = 0.05 \alpha=0.05 α=0.05,求 X ∼ χ 2 ( 18 ) X \sim \chi^2(18) X∼χ2(18) 单侧下限分位点。
(3)当 α = 0.025 \alpha=0.025 α=0.025,求 X ∼ χ 2 ( 28 ) X \sim \chi^2(28) X∼χ2(28) 单侧上限分位点。
解:
(1)当 α = 0.01 \alpha=0.01 α=0.01, χ 0.995 2 ( 15 ) = 4.601 , χ 0.005 2 ( 15 ) = 32.801 \chi^2_{0.995}(15)=4.601, \quad \chi^2_{0.005}(15)=32.801 χ0.9952(15)=4.601,χ0.0052(15)=32.801
(2)当 α = 0.05 \alpha=0.05 α=0.05, χ 0.95 2 ( 18 ) = 9.390 \chi^2_{0.95}(18)=9.390 χ0.952(18)=9.390
(3)当 α = 0.025 \alpha=0.025 α=0.025, χ 0.025 2 ( 28 ) = 44.461 \chi^2_{0.025}(28)=44.461 χ0.0252(28)=44.461
例3:
X 1 , ⋯ , X n X_1, \cdots, X_n X1,⋯,Xn 相互独立, X i ∼ N ( μ , σ 2 ) , i = 1 , 2 , ⋯ , n X_i \sim N(\mu, \sigma^2), \; i=1, 2, \cdots, n Xi∼N(μ,σ2),i=1,2,⋯,n, Y = ∑ i = 1 n ( X i − μ σ ) 2 Y = \sum_{i=1}^n(\frac{X_i - \mu}{\sigma})^2 Y=∑i=1n(σXi−μ)2,求 Y Y Y 的分布。
解:
X i ∼ N ( μ , σ 2 ) , i = 1 , 2 , ⋯ , n X_i \sim N(\mu, \sigma^2), \quad i=1, 2, \cdots, n Xi∼N(μ,σ2),i=1,2,⋯,n令
Y i = X i − μ σ ∼ N ( 0 , 1 ) , i = 1 , 2 , ⋯ , n ⇒ Y i 2 = ( X i − μ σ ) 2 ∼ χ 2 ( 1 ) ; ⇒ Y = ∑ i = 1 n Y i 2 = ∑ i = 1 n ( X i − μ σ ) 2 ∼ χ 2 ( n ) \begin{aligned}&Y_i = \frac{X_i-\mu}{\sigma} \sim N(0, 1), \qquad i=1, 2, \cdots, n \\ \Rightarrow &Y_i^2=(\frac{X_i - \mu}{\sigma})^2 \sim \chi^2(1); \\ \Rightarrow &Y = \sum_{i=1}^nY_i^2=\sum_{i=1}^n(\frac{X_i-\mu}{\sigma})^2 \sim \chi^2(n)\end{aligned} ⇒⇒Yi=σXi−μ∼N(0,1),i=1,2,⋯,nYi2=(σXi−μ)2∼χ2(1);Y=i=1∑nYi2=i=1∑n(σXi−μ)2∼χ2(n)
2.4 t ( n ) t(n) t(n) 分布
- 定义:
X ∼ N ( 0 , 1 ) X \sim N(0, 1) X∼N(0,1), Y ∼ χ 2 ( n ) Y \sim \chi^2(n) Y∼χ2(n), X X X 与 Y Y Y 独立,则称
t = X Y / n t = \frac{X}{\sqrt{Y / n}} t=Y/nX服从自由度为 n n n 的 t t t分布,记为 t ∼ t ( n ) t \sim t(n) t∼t(n)。 - 性质:
- 关于Y轴对称。
- 当 n → ∞ n \rightarrow \infty n→∞ 时, t ( n ) = N ( 0 , 1 ) t(n) = N(0, 1) t(n)=N(0,1)( n > 45 n>45 n>45 时基本相同)
- 分位点: (以
α
=
0.05
\alpha=0.05
α=0.05 为例)
- 双侧分位点:
t
0.975
(
n
)
=
−
t
0.025
t_{0.975}(n)=-t_{0.025}
t0.975(n)=−t0.025 和
t
0.025
(
n
)
t_{0.025}(n)
t0.025(n)
- 非小概率事件区间: { − t 0.025 ( n ) , t 0.025 ( n ) } \{ -t_{0.025}(n), t_{0.025}(n) \} {−t0.025(n),t0.025(n)}
- 小概率事件区间: { − ∞ , − t 0.025 ( n ) } ∪ { t 0.025 ( n ) , + ∞ } \{ -\infty, -t_{0.025}(n) \} \cup \{ t_{0.025}(n), +\infty \} {−∞,−t0.025(n)}∪{t0.025(n),+∞}
- 单侧上限分位点 (过高异常):
t
0.05
(
n
)
t_{0.05}(n)
t0.05(n)
- 非小概率事件区间: { − ∞ , t 0.05 ( n ) } \{ -\infty, t_{0.05}(n) \} {−∞,t0.05(n)}
- 小概率事件区间: { t 0.05 ( n ) , + ∞ } \{ t_{0.05}(n), +\infty \} {t0.05(n),+∞}
- 单侧下限分位点 (过低异常):
t
0.95
(
n
)
=
−
t
0.05
(
n
)
t_{0.95}(n) = -t_{0.05}(n)
t0.95(n)=−t0.05(n)
- 非小概率事件区间: { − t 0.05 ( n ) , + ∞ } \{ -t_{0.05}(n), +\infty \} {−t0.05(n),+∞}
- 小概率事件区间: { − ∞ , − t 0.05 ( n ) } \{ -\infty, -t_{0.05}(n) \} {−∞,−t0.05(n)}
- 双侧分位点:
t
0.975
(
n
)
=
−
t
0.025
t_{0.975}(n)=-t_{0.025}
t0.975(n)=−t0.025 和
t
0.025
(
n
)
t_{0.025}(n)
t0.025(n)
例4:
(1)当 α = 0.05 \alpha=0.05 α=0.05,求 t ∼ t ( 17 ) t \sim t(17) t∼t(17) 双侧分位点。
(2)当 α = 0.01 \alpha=0.01 α=0.01,求 t ∼ t ( 12 ) t \sim t(12) t∼t(12) 单侧上限分位点。
(3)当 α = 0.1 \alpha=0.1 α=0.1,求 t ∼ t ( 29 ) t \sim t(29) t∼t(29) 单侧下限分位点。
解:
(1)当 α = 0.05 \alpha=0.05 α=0.05 时, t ∼ t ( 17 ) t \sim t(17) t∼t(17) 的双侧分位点为 ± t 0.025 ( 17 ) = ± 2.1098 \pm t_{0.025}(17) = \pm 2.1098 ±t0.025(17)=±2.1098
(2)当 α = 0.01 \alpha=0.01 α=0.01 时, t ∼ t ( 12 ) t \sim t(12) t∼t(12) 的单侧上限分位点为 t 0.01 ( 12 ) = 2.6810 t_{0.01}(12) = 2.6810 t0.01(12)=2.6810
(3)当 α = 0.1 \alpha=0.1 α=0.1 时, t ∼ t ( 29 ) t \sim t(29) t∼t(29) 的单侧下限分位点为 t 0.9 ( 29 ) = − t 0.1 ( 29 ) = − 1.3114 t_{0.9} (29)= -t_{0.1}(29) = -1.3114 t0.9(29)=−t0.1(29)=−1.3114
例5:
X , Y , Z X, Y, Z X,Y,Z 相互独立,均服从 N ( 0 , 1 ) N(0, 1) N(0,1) 分布,则:
X 2 + Y 2 + Z 2 ∼ χ 2 ( 3 ) X^2+Y^2+Z^2 \sim \chi^2(3) X2+Y2+Z2∼χ2(3) X X 2 + Y 2 2 ∼ t ( 2 ) \frac{X}{\sqrt{\frac{X^2 + Y^2}{2}}} \sim t(2) 2X2+Y2X∼t(2)
2.5 F ( n , m ) F(n, m) F(n,m) 分布
- 定义:
U ∼ χ 2 ( n 1 ) U \sim \chi^2(n_1) U∼χ2(n1), V ∼ χ 2 ( n 2 ) V \sim \chi^2(n_2) V∼χ2(n2), U U U 与 V V V 独立,则称:
F = U / n 1 V / n 2 F=\frac{U / n_1}{V / n_2} F=V/n2U/n1服从自由度为 ( n 1 , n 2 ) (n_1, n_2) (n1,n2) 的 F F F分布,记为 F ( n 1 , n 2 ) F(n_1, n_2) F(n1,n2)。 - 性质:
- 若 X ∼ F ( n 1 , n 2 ) X \sim F(n_1, n_2) X∼F(n1,n2),则 1 X ∼ F ( n 2 , n 1 ) \frac1{X} \sim F(n_2, n_1) X1∼F(n2,n1)
- F 1 − α ( n 1 , n 2 ) = 1 F α ( n 2 , n 1 ) F_{1-\alpha}(n_1, n_2) = \frac{1}{F_{\alpha}(n_2, n_1)} F1−α(n1,n2)=Fα(n2,n1)1
- 若 t ∼ t ( n ) t \sim t(n) t∼t(n),则 t 2 = ( X Y / n ) 2 = X 2 / 1 Y / n ∼ F ( 1 , n ) t^2 = (\frac{X}{\sqrt{Y / n}})^2 = \frac{X^2 / 1}{Y / n} \sim F(1, n) t2=(Y/nX)2=Y/nX2/1∼F(1,n)
- 分位点: (以
α
=
0.05
\alpha=0.05
α=0.05 为例)
- 双侧分位点:
F
0.975
(
n
,
m
)
=
1
F
0.025
(
m
,
n
)
F_{0.975}(n, m) = \frac{1}{F_{0.025}(m, n)}
F0.975(n,m)=F0.025(m,n)1 及
F
0.025
(
n
,
m
)
F_{0.025}(n, m)
F0.025(n,m)
- 非小概率事件区间: { 1 F 0.025 ( m , n ) , F 0.025 ( n , m ) } \{ \frac 1{F_{0.025}(m, n)}, F_{0.025}(n, m) \} {F0.025(m,n)1,F0.025(n,m)}
- 小概率事件区间: { 0 , 1 F 0.025 ( m , n ) } ∪ { F 0.025 ( n , m ) , + ∞ } \{ 0, \frac{1}{F_{0.025}(m, n)} \} \cup \{ F_{0.025}(n, m), +\infty \} {0,F0.025(m,n)1}∪{F0.025(n,m),+∞}
- 单侧上限分位点 (过高异常):
F
0.05
(
n
,
m
)
F_{0.05}(n, m)
F0.05(n,m)
- 非小概率事件区间: { 0 , F 0.05 ( n , m ) } \{ 0, F_{0.05}(n, m) \} {0,F0.05(n,m)}
- 小概率事件区间: { F 0.05 ( n , m ) , + ∞ } \{ F_{0.05}(n, m), +\infty \} {F0.05(n,m),+∞}
- 单侧下限分位点 (过低异常):
F
0.95
(
n
,
m
)
=
1
F
0.05
(
m
,
n
)
F_{0.95}(n, m) = \frac 1{F_{0.05}(m, n)}
F0.95(n,m)=F0.05(m,n)1
- 非小概率事件区间: { 1 F 0.05 ( m , n ) , + ∞ } \{ \frac 1{F_{0.05}(m, n)}, +\infty \} {F0.05(m,n)1,+∞}
- 小概率事件区间: { 0 , 1 F 0.05 ( m , n ) } \{ 0, \frac 1{F_{0.05}(m, n)} \} {0,F0.05(m,n)1}
- 双侧分位点:
F
0.975
(
n
,
m
)
=
1
F
0.025
(
m
,
n
)
F_{0.975}(n, m) = \frac{1}{F_{0.025}(m, n)}
F0.975(n,m)=F0.025(m,n)1 及
F
0.025
(
n
,
m
)
F_{0.025}(n, m)
F0.025(n,m)
例6:
(1)当 α = 0.05 \alpha = 0.05 α=0.05,求 X ∼ F ( 9 , 15 ) X \sim F(9, 15) X∼F(9,15) 双侧分位点。
(2)当 α = 0.001 \alpha = 0.001 α=0.001,求 X ∼ F ( 13 , 11 ) X \sim F(13, 11) X∼F(13,11) 单侧下限分位点。
(3)当 α = 0.025 \alpha = 0.025 α=0.025,求 X ∼ F ( 25 , 17 ) X \sim F(25, 17) X∼F(25,17) 单侧上限分位点。
解:
(1)当 α = 0.05 \alpha = 0.05 α=0.05, X ∼ F ( 9 , 15 ) X \sim F(9, 15) X∼F(9,15)时,双侧分位点为:
F 0.975 ( 9 , 15 ) = 1 F 0.025 ( 15 , 9 ) = 1 3.77 F 0.025 ( 9 , 15 ) = 3.12 \begin{aligned} &F_{0.975}(9, 15) = \frac 1{F_{0.025}(15, 9)} = \frac 1{3.77} \\ &F_{0.025}(9, 15) = 3.12 \end{aligned} F0.975(9,15)=F0.025(15,9)1=3.771F0.025(9,15)=3.12(2)当 α = 0.001 \alpha = 0.001 α=0.001, X ∼ F ( 13 , 11 ) X \sim F(13, 11) X∼F(13,11) 单侧下限分位点为:
F 0.99 ( 13 , 11 ) = 1 F 0.01 ( 11 , 13 ) = 2 4.1 + 3.96 F_{0.99}(13, 11) = \frac 1{F_{0.01}(11, 13)} = \frac 2{4.1 + 3.96} F0.99(13,11)=F0.01(11,13)1=4.1+3.962(3)当 α = 0.025 \alpha = 0.025 α=0.025, X ∼ F ( 25 , 17 ) X \sim F(25, 17) X∼F(25,17) 单侧上限分位点为:
F 0.025 ( 25 , 17 ) = 2.56 F_{0.025}(25, 17) = 2.56 F0.025(25,17)=2.56
例7:
总体 X ∼ N ( 0 , 1 ) X \sim N(0, 1) X∼N(0,1), X 1 , X 2 , ⋯ , X n X_1, X_2, \cdots, X_n X1,X2,⋯,Xn 是样本,下列统计量各服从什么分布?
(1) X 1 2 + X 2 2 X_1^2 + X_2^2 \quad X12+X22(2) X 1 − X 2 X 3 2 + X 4 2 \frac{X_1-X_2}{\sqrt{X_3^2+X_4^2}} \quad X32+X42X1−X2(3) X 2 ∣ X 6 ∣ \frac{X_2}{|X_6|} \quad ∣X6∣X2(4) ( n 3 − 1 ) ∑ i = 1 3 X i 2 ∑ i = 4 n X i 2 \frac{(\frac n3 - 1)\sum_{i=1}^3X_i^2}{\sum_{i=4}^nX_i^2} ∑i=4nXi2(3n−1)∑i=13Xi2
解:
(1) X 1 2 + X 2 2 ∼ χ 2 ( 2 ) X_1^2 + X_2^2 \sim \chi^2(2) X12+X22∼χ2(2)
(2) X 1 − X 2 ∼ N ( 0 , 2 ) X 1 − X 2 2 ∼ N ( 0 , 1 ) X 3 2 + X 4 2 ∼ χ 2 ( 2 ) \begin{aligned}X_1-X_2 &\sim N(0, 2) \\ \qquad \frac{X_1-X_2}{\sqrt{2}} &\sim N(0, 1) \\ X_3^2+X_4^2 &\sim \chi^2(2) \end{aligned} X1−X22X1−X2X32+X42∼N(0,2)∼N(0,1)∼χ2(2) X 1 , ⋯ , X n \qquad X_1, \cdots, X_n X1,⋯,Xn 相互独立,则:
X 1 − X 2 X 3 2 + X 4 2 = X 1 − X 2 2 X 3 2 + X 4 2 2 ∼ t ( 2 ) \frac{X_1-X_2}{\sqrt{X_3^2+X_4^2}}=\frac{\frac{X_1-X_2}{\sqrt 2}}{\sqrt{\frac{X_3^2+X_4^2}{2}}} \sim t(2) X32+X42X1−X2=2X32+X422X1−X2∼t(2)
(3) X 2 ∼ N ( 0 , 1 ) X 6 2 ∼ χ 2 ( 1 ) X_2 \sim N(0, 1) \qquad X_6^2 \sim \chi^2(1) X2∼N(0,1)X62∼χ2(1) X 2 ∣ X 6 ∣ = X 2 X 6 2 1 ∼ t ( 1 ) \frac{X_2}{|X_6|} = \frac{X_2}{\sqrt \frac {X_6^2}1} \sim t(1) ∣X6∣X2=1X62X2∼t(1)
(4) ∑ i = 1 3 X i 2 ∼ χ 2 ( 3 ) , ∑ i = 4 n X i 2 ∼ χ 2 ( n − 3 ) \sum_{i=1}^3X_i^2 \sim \chi^2(3), \qquad \sum_{i=4}^nX_i^2 \sim \chi^2(n-3) i=1∑3Xi2∼χ2(3),i=4∑nXi2∼χ2(n−3) ( n 3 − 1 ) ∑ i = 1 3 X i 2 ∑ i = 4 n X i 2 = ∑ i = 1 3 X i 2 3 ∑ i = 4 n X i 2 n − 3 ∼ F ( 3 , n − 3 ) \begin{aligned} \frac{(\frac n3 - 1)\sum_{i=1}^3X_i^2}{\sum_{i=4}^nX_i^2} = \frac{\frac{\sum_{i=1}^3X_i^2}{3}}{\frac{\sum_{i=4}^nX_i^2}{n-3}} \sim F(3, n-3) \end{aligned} ∑i=4nXi2(3n−1)∑i=13Xi2=n−3∑i=4nXi23∑i=13Xi2∼F(3,n−3)
三、正态总体的抽样分布
3.1 定理一: X ˉ − μ σ / n ∼ N ( 0 , 1 ) \frac{\bar{X}-\mu}{\sigma/\sqrt{n}}\sim N(0, 1) σ/nXˉ−μ∼N(0,1)( σ \sigma σ已知)
3.1.1 μ ⇐ X ˉ \mu \Leftarrow \bar{X} μ⇐Xˉ 分布
总体
X
∼
N
(
μ
,
σ
2
)
X\sim N(\mu, \sigma^2)
X∼N(μ,σ2),
X
1
,
X
2
,
⋯
,
X
n
X_1, X_2, \cdots, X_n
X1,X2,⋯,Xn 为样本,构造样本均数
X
ˉ
n
=
1
n
∑
i
=
1
n
X
i
\bar{X}_n=\frac 1n\sum_{i=1}^nX_i
Xˉn=n1∑i=1nXi,则:
X
ˉ
−
μ
σ
/
n
∼
N
(
0
,
1
)
(
σ
已知)
\frac{\bar{X}-\mu}{\sigma/\sqrt{n}}\sim N(0, 1) \qquad {\text{(}}\sigma {\text{已知)}}
σ/nXˉ−μ∼N(0,1)(σ已知)
证明:
X ∼ N ( μ , σ 2 ) X\sim N(\mu, \sigma^2) X∼N(μ,σ2), X 1 , X 2 , ⋯ , X n X_1, X_2, \cdots, X_n X1,X2,⋯,Xn 相互独立,由正态分布的可加性有:
∑ i = 1 n X i ∼ N ( n μ , n σ 2 ) ⟹ 1 n ∑ i = 1 n X i ∼ N ( μ , σ 2 n ) ⟹ X ˉ n − μ σ / n ∼ N ( 0 , 1 ) \sum_{i=1}^nX_i\sim N(n\mu,n\sigma^2) \qquad \Longrightarrow \qquad \frac 1n \sum_{i=1}^nX_i\sim N(\mu, \frac{\sigma^2}n) \qquad \Longrightarrow \qquad \frac{\bar{X}_n-\mu}{\sigma/\sqrt{n}}\sim N(0, 1) i=1∑nXi∼N(nμ,nσ2)⟹n1i=1∑nXi∼N(μ,nσ2)⟹σ/nXˉn−μ∼N(0,1)
3.1.2 p ⇐ k / n p\Leftarrow k/n p⇐k/n 分布
总体
X
∼
B
(
1
,
p
)
X\sim B(1, p)
X∼B(1,p),
X
1
,
X
2
,
⋯
,
X
n
X_1, X_2, \cdots, X_n
X1,X2,⋯,Xn 为样本,构造样本均数
X
ˉ
n
=
1
n
∑
i
=
1
n
X
i
=
k
n
\bar{X}_n=\frac 1n \sum_{i=1}^nX_i=\frac kn
Xˉn=n1∑i=1nXi=nk
⇒
k
/
n
−
p
p
(
1
−
p
)
/
n
∼
N
(
0
,
1
)
\Rightarrow \frac{k/n - p}{\sqrt{p(1-p)/n}}\sim N(0, 1)
⇒p(1−p)/nk/n−p∼N(0,1)
证明:
X ∼ B ( 1 , p ) X\sim B(1, p) X∼B(1,p), X 1 , X 2 , ⋯ , X n X_1, X_2, \cdots, X_n X1,X2,⋯,Xn 相互独立,由中心极限定理得:
∑ i = 1 n X i ∼ N { n p , n p ( 1 − p ) } ⇒ 1 n ∑ i = 1 n X i = k n ∼ N ( p , p ( 1 − p ) n ) ⇒ k / n − p p ( 1 − p ) / n ∼ N ( 0 , 1 ) \sum_{i=1}^nX_i\sim N\{np, np(1-p)\} \\ \begin{aligned}&\Rightarrow \frac 1n \sum_{i=1}^nX_i=\frac kn \sim N(p, \frac{p(1-p)}{n}) \\ &\Rightarrow \frac{k/n-p}{\sqrt{p(1-p)/n}}\sim N(0, 1)\end{aligned} i=1∑nXi∼N{np,np(1−p)}⇒n1i=1∑nXi=nk∼N(p,np(1−p))⇒p(1−p)/nk/n−p∼N(0,1)
3.2 定理二: ( n − 1 ) S 2 σ 2 ∼ χ 2 ( n − 1 ) \frac{(n-1)S^2}{\sigma^2}\sim\chi^2(n-1) σ2(n−1)S2∼χ2(n−1)
σ 2 ⇐ S 2 \sigma^2\Leftarrow S^2 σ2⇐S2 分布
总体
X
∼
N
(
μ
,
σ
2
)
X\sim N(\mu, \sigma^2)
X∼N(μ,σ2),
X
1
,
X
2
,
⋯
,
X
n
X_1, X_2, \cdots, X_n
X1,X2,⋯,Xn 为样本,构造样本均数
X
ˉ
n
=
1
n
∑
i
=
1
n
X
i
\bar{X}_n=\frac 1n \sum_{i=1}^nX_i
Xˉn=n1∑i=1nXi,样本方差
S
2
=
1
n
−
1
∑
i
=
1
n
(
X
i
−
X
ˉ
n
)
2
S^2=\frac 1{n-1}\sum_{i=1}^n(X_i-\bar{X}_n)^2
S2=n−11∑i=1n(Xi−Xˉn)2,则有:
(1)
X
ˉ
\bar{X}
Xˉ 与
S
2
S^2
S2 独立;
(2)
(
n
−
1
)
S
2
σ
2
∼
χ
2
(
n
−
1
)
\frac{(n-1)S^2}{\sigma^2}\sim \chi^2(n-1)
σ2(n−1)S2∼χ2(n−1)
3.3 定理三: X ˉ − μ S / n ∼ t ( n ) \frac{\bar{X}-\mu}{S/\sqrt{n}}\sim t(n) S/nXˉ−μ∼t(n)( σ \sigma σ未知)
μ ⇐ X ˉ \mu \Leftarrow \bar{X} μ⇐Xˉ分布
总体
X
∼
N
(
μ
,
σ
2
)
X\sim N(\mu, \sigma^2)
X∼N(μ,σ2),
X
1
,
X
2
,
⋯
,
X
n
X_1, X_2, \cdots, X_n
X1,X2,⋯,Xn 为样本,构造样本均数
X
ˉ
n
=
1
n
∑
i
=
1
n
X
i
\bar{X}_n = \frac 1n\sum_{i=1}^nX_i
Xˉn=n1∑i=1nXi,样本方差
S
2
=
1
n
−
1
∑
i
=
1
n
(
X
i
−
X
ˉ
n
)
2
S^2=\frac 1{n-1}\sum_{i=1}^n(X_i-\bar{X}_n)^2
S2=n−11∑i=1n(Xi−Xˉn)2,则有:
t
=
X
ˉ
n
−
μ
S
/
n
∼
t
(
n
−
1
)
t=\frac{\bar{X}_n-\mu}{S/\sqrt{n}}\sim t(n-1)
t=S/nXˉn−μ∼t(n−1)
证明:
X ∼ N ( 0 , 1 ) Y ∼ χ 2 ( n ) X 与 Y 独立 } ⟹ X Y n ∼ t ( n ) \begin{aligned}&X\sim N(0, 1) \\ &Y\sim \chi^2(n) \\ &X\text{与}Y\text{独立} \end{aligned}\Bigg\} \qquad \Longrightarrow \qquad \frac X{\sqrt{\frac Yn}} \sim t(n) X∼N(0,1)Y∼χ2(n)X与Y独立}⟹nYX∼t(n) 总体 X ∼ N ( μ , σ 2 ) X\sim N(\mu, \sigma^2) X∼N(μ,σ2), X 1 , X 2 , ⋯ , X n X_1, X_2, \cdots, X_n X1,X2,⋯,Xn 为样本,样本均数 X ˉ n = 1 n ∑ i = 1 n X i \bar{X}_n=\frac 1n \sum_{i=1}^nX_i Xˉn=n1∑i=1nXi
X ˉ n ∼ N ( μ , σ 2 n ) ⟹ X ˉ n − μ σ / n ∼ N ( 0 , 1 ) \bar{X}_n \sim N(\mu, \frac{\sigma^2}n) \qquad \Longrightarrow \qquad \frac{\bar{X}_n-\mu}{\sigma/\sqrt{n}} \sim N(0, 1) Xˉn∼N(μ,nσ2)⟹σ/nXˉn−μ∼N(0,1) ( n − 1 ) S 2 σ 2 ∼ χ 2 ( n − 1 ) \frac{(n-1)S^2}{\sigma^2}\sim \chi^2(n-1) σ2(n−1)S2∼χ2(n−1) 且 X ˉ \bar{X} Xˉ 与 S 2 S^2 S2 相互独立,则:
t = X ˉ n − μ σ / n ( n − 1 ) S 2 σ 2 1 n − 1 = X ˉ − μ S / n ∼ t ( n − 1 ) t=\frac{\frac{\bar{X}_n-\mu}{\sigma / \sqrt n}}{\sqrt{\frac{(n-1)S^2}{\sigma^2}\frac 1{n-1}}} =\frac{\bar{X}-\mu}{S/\sqrt{n}}\sim t(n-1) t=σ2(n−1)S2n−11σ/nXˉn−μ=S/nXˉ−μ∼t(n−1)
3.4 定理四: S 1 2 / S 2 2 σ 1 2 / σ 2 2 ∼ F ( n 1 , n 2 ) \frac{S_1^2/S_2^2}{\sigma_1^2/\sigma_2^2}\sim F(n_1, n_2) σ12/σ22S12/S22∼F(n1,n2)
μ 1 − μ 2 ⇐ X ˉ − Y ˉ \mu_1-\mu_2 \Leftarrow \bar{X}-\bar{Y} μ1−μ2⇐Xˉ−Yˉ分布; σ 1 2 / σ 2 2 ⇐ S 1 2 / S 2 2 \sigma_1^2/\sigma_2^2\Leftarrow S_1^2/S_2^2 σ12/σ22⇐S12/S22分布
总体
X
∼
N
(
μ
1
,
σ
1
2
)
X\sim N(\mu_1, \sigma_1^2)
X∼N(μ1,σ12),样本
X
1
,
X
2
,
⋯
,
X
n
X_1, X_2, \cdots, X_n
X1,X2,⋯,Xn,样本均数
X
ˉ
n
\bar{X}_n
Xˉn,样本方差
S
1
2
S_1^2
S12
总体
Y
∼
N
(
μ
2
,
σ
2
2
)
Y\sim N(\mu_2, \sigma_2^2)
Y∼N(μ2,σ22),样本
Y
1
,
Y
2
,
⋯
,
Y
m
Y_1, Y_2, \cdots, Y_m
Y1,Y2,⋯,Ym,样本均数
Y
ˉ
m
\bar{Y}_m
Yˉm,样本方差
S
2
2
S_2^2
S22
且
X
X
X 与
Y
Y
Y 独立,则有:
(1)
(
X
ˉ
n
−
Y
ˉ
m
)
−
(
μ
1
−
μ
2
)
σ
1
2
/
n
+
σ
2
2
/
m
∼
N
(
0
,
1
)
(
σ
1
2
,
σ
2
2
已知)
\frac{(\bar{X}_n-\bar{Y}_m)-(\mu_1-\mu_2)}{\sqrt{\sigma_1^2/n + \sigma_2^2/m}}\sim N(0, 1) \qquad \text{(} \sigma_1^2, \sigma_2^2 \text{已知)}
σ12/n+σ22/m(Xˉn−Yˉm)−(μ1−μ2)∼N(0,1)(σ12,σ22已知)(2)
(
X
ˉ
n
−
Y
ˉ
m
)
−
(
μ
1
−
μ
2
)
S
w
1
n
+
1
m
∼
t
(
n
+
m
−
2
)
(
σ
1
2
=
σ
2
2
未知)
\frac{(\bar{X}_n-\bar{Y}_m)-(\mu_1-\mu_2)}{S_w\sqrt{\frac 1n+\frac 1m}} \sim t(n+m-2) \qquad \text{(} \sigma_1^2 = \sigma_2^2 \text{未知)}
Swn1+m1(Xˉn−Yˉm)−(μ1−μ2)∼t(n+m−2)(σ12=σ22未知)
⟨
S
w
2
=
(
n
−
1
)
S
1
2
+
(
m
−
1
)
S
2
2
n
+
m
−
2
⟩
\Bigg\langle S_w^2=\frac{(n-1)S_1^2+(m-1)S_2^2}{n+m-2} \Bigg\rangle
⟨Sw2=n+m−2(n−1)S12+(m−1)S22⟩(3)
S
1
2
/
S
2
2
σ
1
2
/
σ
2
2
∼
F
(
n
−
1
,
m
−
1
)
\frac{S_1^2/S_2^2}{\sigma_1^2/\sigma_2^2}\sim F(n-1, m-1)
σ12/σ22S12/S22∼F(n−1,m−1)
3.5 定理汇总
参数 | 分布 |
---|---|
μ \mu μ | X ˉ n − μ σ / n ∼ N ( 0 , 1 ) ( σ 已知) \frac{\bar{X}_n-\mu}{\sigma/\sqrt{n}}\sim N(0, 1) \qquad \text{(}\sigma\text{已知)} σ/nXˉn−μ∼N(0,1)(σ已知) X ˉ n − μ S / n ∼ t ( n − 1 ) ( σ 未知) \frac{\bar{X}_n-\mu}{S/\sqrt{n}}\sim t(n-1) \qquad \text{(}\sigma\text{未知)} S/nXˉn−μ∼t(n−1)(σ未知) |
σ 2 \sigma^2 σ2 | ( n − 1 ) S 2 σ 2 ∼ χ 2 ( n − 1 ) \frac{(n-1)S^2}{\sigma^2}\sim \chi^2(n-1) σ2(n−1)S2∼χ2(n−1) |
σ 1 2 / σ 2 2 \sigma_1^2/\sigma_2^2 σ12/σ22 | S 1 2 / S 2 2 σ 1 2 / σ 2 2 ∼ F ( n − 1 , m − 1 ) \frac{S_1^2/S_2^2}{\sigma_1^2/\sigma_2^2}\sim F(n-1, m-1) σ12/σ22S12/S22∼F(n−1,m−1) |
p p p | k / n − p p ( 1 − p ) / n ∼ N ( 0 , 1 ) \frac{k/n-p}{\sqrt{p(1-p)/n}}\sim N(0, 1) p(1−p)/nk/n−p∼N(0,1) |
μ 1 − μ 2 \mu_1-\mu_2 μ1−μ2 | ( X ˉ n − Y ˉ m ) − ( μ 1 − μ 2 ) σ 1 2 / n + σ 2 2 / m ∼ N ( 0 , 1 ) ( σ 1 2 , σ 2 2 已知) \frac{(\bar X_n-\bar Y_m)-(\mu_1-\mu_2)}{\sqrt{\sigma_1^2/n+\sigma_2^2/m}}\sim N(0, 1) \qquad \text{(}\sigma_1^2, \sigma_2^2\text{已知)} σ12/n+σ22/m(Xˉn−Yˉm)−(μ1−μ2)∼N(0,1)(σ12,σ22已知) ( X ˉ n − Y ˉ m ) − ( μ 1 − μ 2 ) S w 1 / n + 1 / m ∼ t ( n + m − 2 ) ( σ 1 2 = σ 2 2 未知) \frac{(\bar X_n-\bar Y_m)-(\mu_1-\mu_2)}{S_w\sqrt{1/n+1/m}}\sim t(n+m-2) \quad \text{(}\sigma_1^2=\sigma_2^2\text{未知)} Sw1/n+1/m(Xˉn−Yˉm)−(μ1−μ2)∼t(n+m−2)(σ12=σ22未知) |