跳转至

单样本均值置信区间

正态分布

当总体方差 \(\sigma^2\) 已知时,可使用正态分布构建置信区间。

\[ Z = \frac{\bar{X} - \mu}{\sigma/\sqrt{n}} \sim N(0, 1) \]
\[ \text{Confidence Interval} = \left(\bar{X}-\frac{z_{1-\alpha/2}\sigma}{\sqrt{n}}, \ \bar{X}+\frac{z_{1-\alpha/2}\sigma}{\sqrt{n}}\right) \]

设均值到置信限的距离为 \(d\),则:

\[ d = \frac{z_{1-\alpha/2}\sigma}{\sqrt{n}} \]
\[ n = \frac{z_{1-\alpha/2}^2\sigma^2}{d^2} \]
\[ \sigma = \frac{d\sqrt{n}}{z_{1-\alpha/2}} \]
\[ \text{Confidence Interval} = \left(-\infty, \ \bar{X}+\frac{z_{1-\alpha}\sigma}{\sqrt{n}}\right) \]

设均值到置信限的距离为 \(d\),则:

\[ d = \frac{z_{1-\alpha}\sigma}{\sqrt{n}} \]
\[ n = \frac{z_{1-\alpha}^2\sigma^2}{d^2} \]
\[ \sigma = \frac{d\sqrt{n}}{z_{1-\alpha}} \]
\[ \text{Confidence Interval} = \left(\bar{X}-\frac{z_{1-\alpha}\sigma}{\sqrt{n}}, \ +\infty\right) \]

设均值到置信限的距离为 \(d\),则:

\[ d = \frac{z_{1-\alpha}\sigma}{\sqrt{n}} \]
\[ n = \frac{z_{1-\alpha}^2\sigma^2}{d^2} \]
\[ \sigma = \frac{d\sqrt{n}}{z_{1-\alpha}} \]

若已知标准差 \(\sigma = 1\) 时,均值到置信限的距离为 \(d\),则当均值到置信限的距离为 \(d'\) 时,标准差 \(\sigma' = d'/d\), 利用此关系可以简化 solve_std 函数的实现,而不必使用 brentq 进行反解。

t 分布

当总体方差 \(\sigma^2\) 未知时,可使用 \(t\) 分布构建置信区间。

\[ T = \frac{\bar{X} - \mu}{S/\sqrt{n}} \sim t(n - 1) \]
\[ \text{Confidence Interval} = \left(\bar{X}-\frac{t_{1-\alpha/2,\ n-1}S}{\sqrt{n}}, \ \bar{X}+\frac{t_{1-\alpha/2,\ n-1}S}{\sqrt{n}}\right) \]

设均值到置信限的距离为 \(d\),则:

\[ d = \frac{t_{1-\alpha/2, \ n-1}S}{\sqrt{n}} \]
\[ S = \frac{d\sqrt{n}}{t_{1-\alpha/2, \ n-1}} \]
\[ \text{Confidence Interval} = \left(-\infty, \ \bar{X}+\frac{t_{1-\alpha,\ n-1}S}{\sqrt{n}}\right) \]

设均值到置信限的距离为 \(d\),则:

\[ d = \frac{t_{1-\alpha, \ n-1}S}{\sqrt{n}} \]
\[ S = \frac{d\sqrt{n}}{t_{1-\alpha, \ n-1}} \]
\[ \text{Confidence Interval} = \left(\bar{X}-\frac{t_{1-\alpha,\ n-1}S}{\sqrt{n}}, \ +\infty\right) \]

设均值到置信限的距离为 \(d\),则:

\[ d = \frac{t_{1-\alpha, \ n-1}S}{\sqrt{n}} \]
\[ S = \frac{d\sqrt{n}}{t_{1-\alpha, \ n-1}} \]

若已知标准差 \(S = 1\) 时,均值到置信限的距离为 \(d\),则当均值到置信限的距离为 \(d'\) 时,标准差 \(S' = d'/d\), 利用此关系可以简化 solve_std 函数的实现,而不必使用 brentq 进行反解。