Minitab 控制图


Minitab 控制图官方文档


一、数据

数据:

数据1 数据2 数据3 子组ID
601.4 598 601.6 1
601.6 599.8 600.4 1
598 600 598.4 1
601.4 599.8 600 1
599.4 600 596.8 1
600 600 602.8 2
600.2 598.8 600.8 2
601.2 598.2 603.6 2
598.4 599.4 604.2 2
599 599.6 602.4 2
601.2 599.4 598.4 3
601 599.4 599.6 3
600.8 600 603.4 3
597.6 598.8 600.6 3
601.6 599.2 598.4 3
599.4 599.4 598.2 4
601.2 599.6 602 4
598.4 599 599.4 4
599.2 599.2 599.4 4
598.8 600.6 600.8 4
601.4 598.8 600.8 5
599 598.8 598.6 5
601 599.8 600 5
601.6 599.2 600.4 5
601.4 599.4 600.8 5
601.4 600 600.8 6
598.8 600.2 597.2 6
601.4 600.2 600.4 6
598.4 599.6 599.8 6
601.6 599 596.4 6
598.8 599 600.4 7
601.2 599.8 598.2 7
599.6 600.8 598.6 7
601.2 598.8 599.6 7
598.2 598.2 599 7
598.8 600 598.2 8
597.8 599.2 599.4 8
598.2 599.8 599.4 8
598.2 601.2 600.2 8
598.2 600.4 599 8
601.2 600.2 599.4 9
600 599.6 598 9
598.8 599.6 597.6 9
599.4 599.6 598 9
597.2 600.2 597.6 9
600.8 599.2 601.2 10
600.6 599 599 10
599.6 599.6 600.4 10
599.4 600.4 600.6 10
598 600 599 10
600.8 599 602.2 11
597.8 599.6 599.8 11
599.2 599.4 599.8 11
599.2 599.2 601 11
600.6 597.8 601.6 11
598 600.4 601.6 12
598 599.6 600.2 12
598.8 600 601.8 12
601 600.8 601.2 12
600.8 600.4 597.6 12
598.8 599.4 599.8 13
599.4 599 602.8 13
601 598.4 600 13
598.8 599 599.6 13
599.6 599.6 602.2 13
599 598.8 603.8 14
600.4 599.2 603.6 14
598.4 599.6 601.8 14
602.2 598.6 602 14
601 599.8 603.6 14
601.4 599.6 600.8 15
601 599.2 600.2 15
601.2 599.6 600.4 15
601.4 600.2 600.2 15
601.8 599.8 602.2 15
601.6 599.6 598 16
601 600 598.4 16
600.2 599.6 600.8 16
599 599.2 602.8 16
601.2 598.6 597.6 16
601.2 599.6 601.6 17
601.2 601.2 603.4 17
601.2 599.6 597 17
601 600.2 599.8 17
601 600 597.8 17
601.4 600 602.4 18
601.4 599.4 602.2 18
598.8 599.8 600.6 18
598.8 599.2 596.2 18
598.8 599.6 602.4 18
598.2 599.4 601.4 19
601.8 600 599.2 19
601 600 601.6 19
601.4 599.2 600.4 19
601.4 599.4 598 19
599 599.6 601.2 20
601.4 599.8 604.2 20
601.8 599 600.2 20
601.6 599.6 600 20
601.2 599.4 596.8 20

数据4:

子组ID PH值
1 6.05
2 5.99
3 6.11
4 6.13
5 5.87
6 6.05
7 6.23
8 6.49
9 6.15
10 5.89
11 5.87
12 5.99
13 6.07
14 6.17
15 5.86
16 6.07
17 6.01
18 5.87
19 5.66
20 5.58
21 5.62
22 5.89
23 6.02
24 5.93
25 6.05

数据5:

子组ID \(x_1\) \(x_2\) \(x_3\)
1 1.504 4.075 1.971
2 1.685 4.599 2.26
3 1.529 4.1 1.994
4 1.554 4.19 2.024
5 1.604 4.275 2.063
6 1.664 4.341 2.12
7 1.789 4.981 2.287
8 1.723 4.416 2.166
9 1.831 5.196 2.285
10 1.622 4.353 2.135
11 1.683 4.396 2.154
12 1.598 4.329 2.18
13 1.847 5.168 2.331
14 1.793 4.547 2.184
15 1.886 5.259 2.389
16 1.631 4.338 2.073
17 1.543 4.204 2.151
18 1.665 4.48 2.282
19 1.578 4.349 2.128
20 1.533 4.28 2.039
21 1.674 4.504 2.192
22 1.749 4.371 2.155
23 1.83 5.094 2.436
24 1.813 4.989 2.428
25 1.73 4.396 2.16


二、控制限

  • 子组变量控制图的 n>1

  • w 的取值范围 [2, 100]

1. Xbar-R 控制图的样本均值图

\(UCL\ =\ \mu\ +\ k\frac{\sigma}{\sqrt{n_i}}\)

\(CL\ =\ \mu\ =\ \bar{X}\)

\(LCL\ =\ \mu\ -\ k\frac{\sigma}{\sqrt{n_i}}\)

\(\sigma\ =\ \frac{Rbar}{d_2(n_i)}\)


\(μ = \bar{X}\):过程均值

\(k\):检验 1 的参数,默认为 3

\(n_{i}\):子组 i 的观测值个数

\(Rbar\):子组极差的均值

\(d_2\left(n_i\right)\):与括号中指定的值相对应的无偏常量 \(d_2\)

\(d_{2}(n_i)=3.4873+0.0250141 \times n_i-0.00009823 \times n_i^{2}\ \ \ \ n_i\in[51,100]\)


实例:数据1

\[\begin{aligned} & ptp=[3.6,2.8,4.0,2.8,2.6,...,2.6,3.6,2.8] \\ & Rbar=\frac{3.6+2.8+4.0...+2.6+3.6+2.8}{20}=2.72 \\ & d_2(n_i)=d_2(5)=2.326 \\ & \sigma=\frac{Rbar}{d_2(n_i)}=1.169 \end{aligned} \]


2. Xbar-S 控制图的样本均值图

\(UCL\ =\ \mu\ +\ k\frac{\sigma}{\sqrt{n_i}}\)

\(CL\ =\ \mu\ =\ \bar{X}\)

\(LCL\ =\ \mu\ -\ k\frac{\sigma}{\sqrt{n_i}}\)

\(\sigma\ =\ \frac{Sbar}{C_4(n_i)}\)


\(μ = \bar{X}\):过程均值

\(k\):检验 1 的参数,默认为 3

\(n_{i}\):子组 i 的观测值个数

\(Sbar\):子组标准差的均值

\(C_4(·)\):与括号指定的值相对应的无偏常量 \(C_4\) 值。计算公式为:\(C_4(N)\ =\ \sqrt{\frac{2}{N-1}}\frac{\gamma(\frac{N}{2})}{\gamma(\frac{N-1}{2})}\)


实例:数据1

\[\begin{aligned} & std=[1.596,1.090,1.615,...,1.459,1.140] \\ & Sbar=\frac{1.596+1.090+1.615+...+1.459+1.140}{20}=1.148 \\ & C_4(n_i)=C_4(5)=0.940 \\ & \sigma=\frac{Sbar}{C_4(n_i)}=1.221 \end{aligned} \]


3. Xbar 控制图

\(UCL\ =\ \mu\ +\ k\frac{\sigma}{\sqrt{n_i}}\)

\(CL\ =\ \mu\ =\ \bar{X}\)

\(LCL\ =\ \mu\ -\ k\frac{\sigma}{\sqrt{n_i}}\)

\(\sigma\ =\ \frac{\sqrt{\mu_v}}{C_4(d+1)}\)


\(μ = \bar{X}\):过程均值

\(k\):检验 1 的参数,默认为 3

\(n_{i}\):子组 i 的观测值个数

\(\mu_v\):子组方差的均值

\(C_4(·)\):与括号指定的值相对应的无偏常量 \(C_4\)值。计算公式为:\(C_4(N)\ =\ \sqrt{\frac{2}{N-1}}\frac{\gamma(\frac{N}{2})}{\gamma(\frac{N-1}{2})}\)

\(d\):自由度。计算公式为:\(\sum{(n_i\ -\ 1})\)


实例:数据1

\[\begin{aligned} & var=[2.548,1.188,2.608,..., 2.128, 1.300] \\ & \mu_v=\frac{2.548+1.188+2.608+...+2.128+1.300}{20}=1.503 \\ & C_4(d+1)=C_4(m\times\ (n_i-1)+1)=C_4(20\times\ (5-1)+1)=0.997 \\ & \sigma=\frac{\sqrt\mu_v}{C_4(d+1)}=1.230 \end{aligned} \]


4. R 控制图

\(UCL\ =\ \mu_R\ +\ k\sigma\)

\(CL\ =\ \mu_R\ =\ Rbar\)

\(LCL\ =\ \mu_R\ -\ k\sigma\)

\(\sigma\ =\ \frac{d_3\left(n_i\right)}{d_2\left(n_i\right)}·Rbar\)


\(μ_R = Rbar\):子组极差的均值

\(k\):检验 1 的参数,默认为 3

\(n_{i}\):子组 i 的观测值个数

\(d_2\left(n_i\right)\):与括号中指定的值相对应的无偏常量 \(d_2\)

\(d_3\left(n_i\right)\):与括号中指定的值相对应的无偏常量 \(d_3\)

\(d_{2}(n_i)=3.4873+0.0250141 \times n_i-0.00009823 \times n_i^{2}\ \ \ \ n_i\in[51,100]\)

\(d_{3}(n_i)=0.80818-0.0051871 \times n_i-0.00049243 \times n_i^{3}\ \ \ \ n_i\in[51,100]\)


实例:数据1

\[\begin{aligned} & ptp=[3.6,2.8,4.0,2.8,2.6,...,2.6,3.6,2.8] \\ & Rbar=\frac{3.6+2.8+4.0...+2.6+3.6+2.8}{20}=2.72 \\ & d_2(n_i)=d_2(5)=2.326 \\ & d_3(n_i)=d_2(5)=0.8641 \\ & \sigma=\frac{d_3(n_i)}{d_2(n_i)}\times\ Rbar=1.010 \end{aligned} \]


5. S 控制图

\(UCL\ =\ \mu_S\ +\ k\sigma\)

\(CL\ =\ \mu_S\ =\ Sbar\)

\(LCL\ =\ \mu_S\ -\ k\sigma\)

\(\sigma\ =\ \frac{C_5\left(n_i\right)}{C_4\left(n_i\right)}·Sbar\)


\(μ_S = Sbar\):子组标准差的均值

\(k\):检验 1 的参数,默认为 3

\(n_{i}\):子组 i 的观测值个数

\(C_4(·)\):与括号指定的值相对应的无偏常量 \(C_4\) 值。计算公式为:\(C_4(N)\ =\ \sqrt{\frac{2}{N-1}}\frac{\gamma(\frac{N}{2})}{\gamma(\frac{N-1}{2})}\)

\(C_5(·)\):与括号指定的值相对应的无偏常量 \(C_5\) 值。计算公式为:\(C_5(N)\ =\ \sqrt{1-{C_4(N)}^2}\)


实例:数据1

\[\begin{aligned} & std=[1.596,1.090,1.615,...,1.459,1.140] \\ & Sbar=\frac{1.596+1.090+1.615+...+1.459+1.140}{20}=1.148 \\ & C_4(n_i)=C_4(5)=0.940 \\ & C_5(n_i)=C_5(5)=0.341 \\ & \sigma=\frac{C_5(n_i)}{C_4(n_i)}\times Sbar=0.417 \end{aligned} \]


6. 单值控制图

\(UCL\ =\ \mu\ +\ \frac{k\sigma}{\sqrt{n_i}}\ =\ \bar{X}\ +\ k\frac{\bar{R}}{d_2(w)}\)

\(CL\ =\ \mu\ =\ \bar{X}\)

\(LCL\ =\ \mu\ -\ \frac{k\sigma}{\sqrt{n_i}}\ =\ \bar{X}\ -\ k\frac{\bar{R}}{d_2(w)}\)

\(\sigma\ =\ \frac{\bar{R}}{d_2(w)}\)


\(μ = \bar{X}\):过程均值

\(k\):检验 1 的参数,默认为 3

\(n_i\):子组 i 的观测值个数,取值为 1

\(w\):移动极差长度,默认为 2

\(\bar{R}\):移动极差的均值

\(d_2\left(n_i\right)\):与括号中指定的值相对应的无偏常量 \(d_2\)


实例:数据4

\[\begin{aligned} & ptp=[0.06,0.12,0.02,...,0.09,0.12] \\ & \bar{R}=\frac{0.06+0.12+0.02+...+0.09+0.12}{24}=0.152 \\ & d_2(w)=d_2(2)=1.128 \\ & \sigma=\frac{\bar{R}}{d_2(w)}=0.135 \end{aligned} \]


7. 移动极差控制图

\(UCL\ =\ \bar{R}\ +\ k\sigma_R\ =\ \bar{R}+\ k\frac{d_3(w)}{d_2(w)}\bar{R}\)

\(CL\ =\ \bar{R}\)

\(LCL\ =\ \bar{R}\ -\ k\sigma_R\ =\ \bar{R}-\ k\frac{d_3(w)}{d_2(w)}\bar{R}\)

\(\sigma\ =\ \frac{\bar{R}}{d_2(w)}\)

\(\sigma_R\ =\ d_3(w)\sigma\ =\ \frac{d_3(w)}{d_2(w)}\bar{R}\)


\(k\)​:检验 1 的参数,默认为 3

\(w\):移动极差长度,默认为 2

\(\bar{R}\):移动极差的均值

\(d_2\left(n_i\right)\):与括号中指定的值相对应的无偏常量 \(d_2\)

\(d_3\left(n_i\right)\):与括号中指定的值相对应的无偏常量 \(d_3\)


实例:数据4

\[\begin{aligned} & ptp=[0.06,0.12,0.02,...,0.09,0.12] \\ & \bar{R}=\frac{0.06+0.12+0.02+...+0.09+0.12}{24}=0.152 \\ & d_2(w)=d_2(2)=1.128 \\ & d_3(w)=d_3(2)=0.8525 \\ & \sigma=\frac{\bar{R}}{d_2(w)}=0.135 \\ & \sigma_R=d_3(w)\sigma=\frac{d_3(w)}{d_2(w)}\times \bar{R}=0.115 \end{aligned} \]


8. MA 控制图

\(n_i>1\)时:

\(i<mv\)

\(UCL = \mu\ +\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}\)

\(LCL = \mu\ -\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}\)

\(i≥mv\)

\(UCL= \mu\ +\ k(\frac{\sigma}{mv})\sqrt{\frac{mv}{n_i}}\)

\(LCL= \mu\ -\ k(\frac{\sigma}{mv})\sqrt{\frac{mv}{n_i}}\)

\(\sigma\ =\ \frac{\sqrt{\mu_v}}{C_4(d+1)}\)


\(n_i=1\)

\(i<mv\)

\(UCL = \mu\ +\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}\)

\(LCL = \mu\ -\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}\)

\(i≥mv\)

\(UCL= \mu\ +\ k(\frac{\sigma}{mv})\sqrt{\frac{mv}{n_i}}\)

\(LCL= \mu\ -\ k(\frac{\sigma}{mv})\sqrt{\frac{mv}{n_i}}\)

\(\sigma\ =\ \frac{\bar{R}}{d_2(w)}\)


\(i\)​:遍历子组个数的取值(不是子组的观测值个数

\(mv\):移动均值长度,默认为3​


实例:

数据1:\(n_i>1\)

\[\begin{aligned} \sigma=\frac{\sqrt{\mu_v}}{C_4(d+1)}=1.230 \\ \\ ①\ i<mv: \\ & UCL_1=\mu\ +\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}=600.072+3(\frac{1.230}{1})\sqrt{\frac{1}{5}}=601.722 \\ & UCL_2=\mu\ +\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}=600.072+3(\frac{1.230}{2})\sqrt{\frac{2}{5}}=601.239 \\ ②\ i≥mv: \\ & UCL_4=\mu\ +\ k(\frac{\sigma}{mv})\sqrt{\frac{mv}{n_i}}=600.072+3(\frac{1.230}{3})\sqrt{\frac{3}{5}}=601.024 \\ \end{aligned} \]

数据4:\(n_i=1\)

\[\begin{aligned} \sigma=\frac{\bar{R}}{d_2(w)}=0.135 \\ \\ ①\ i<mv: \\ & UCL_1=\mu\ +\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}=5.985+3(\frac{0.135}{1})\sqrt{\frac{1}{1}}=6.390 \\ & UCL_2=\mu\ +\ k(\frac{\sigma}{i})\sqrt{\frac{i}{n_i}}=5.985+3(\frac{0.135}{2})\sqrt{\frac{2}{1}}=6.272 \\ ②\ i≥mv: \\ & UCL_4=\mu\ +\ k(\frac{\sigma}{mv})\sqrt{\frac{mv}{n_i}}=5.985+3(\frac{0.135}{3})\sqrt{\frac{3}{1}}=6.219 \\ \end{aligned} \]


9. EWMA 控制图

\(n_i>1\) 时:

\(UCL\ =\ \mu\ +\ k\sigma_Z\ = \mu\ +\ k\frac{\sigma}{\sqrt{n_i}}\sqrt{(\frac{r}{2-r})\left[1-\left(1-r\right)^{2i}\right]}\)

\(CL\ =\ \mu\ =\ \bar{X}\)

\(LCL\ =\ \mu\ -\ k\sigma_Z\ = \mu\ -\ k\frac{\sigma}{\sqrt{n_i}}\sqrt{(\frac{r}{2-r})\left[1-\left(1-r\right)^{2i}\right]}\)

\(\sigma\ =\ \frac{\sqrt{\mu_v}}{C_4(d+1)}\)


\(n_i=1\)

\(UCL\ =\ \mu\ +\ k\sigma_Z\ = \mu\ +\ k\frac{\sigma}{\sqrt{n_i}}\sqrt{(\frac{r}{2-r})\left[1-\left(1-r\right)^{2i}\right]}\)

\(CL\ =\ \mu\ =\ \bar{X}\)

\(LCL\ =\ \mu\ -\ k\sigma_Z\ = \mu\ -\ k\frac{\sigma}{\sqrt{n_i}}\sqrt{(\frac{r}{2-r})\left[1-\left(1-r\right)^{2i}\right]}\)

\(\sigma\ =\ \frac{\bar{R}}{d_2(w)}\)


\(r\)​:EWMA权重,默认为0.2


实例:

数据1:\(n_i>1\)

\[\begin{aligned} \sigma=\frac{\sqrt{\mu_v}}{C_4(d+1)}=1.230 \\ & UCL_1=600.072+3\times \frac{1.230}{\sqrt{5}}\sqrt{(\frac{0.2}{2-0.2})[1-(1-0.2)^{2}]}=600.402 \\ & UCL_1=600.072+3\times \frac{1.230}{\sqrt{5}}\sqrt{(\frac{0.2}{2-0.2})[1-(1-0.2)^{4}]}=600.495 \\ \end{aligned} \]

数据4:\(n_i=1\)

\[\begin{aligned} \sigma=\frac{\bar{R}}{d_2(w)}=0.135 \\ & UCL_1=5.985+3\times \frac{0.135}{\sqrt{1}}\sqrt{(\frac{0.2}{2-0.2})[1-(1-0.2)^{2}]}=6.066 \\ & UCL_1=5.985+3\times \frac{0.135}{\sqrt{1}}\sqrt{(\frac{0.2}{2-0.2})[1-(1-0.2)^{4}]}=6.089 \\ \end{aligned} \]


10. T方控制图

子组中的数据:数据1、数据2

\(UCL = \frac{p(m-1)(n-1)}{mn-m-p+1}F(1-p\alpha,\ p,\ mn-m-p+1)\)

\(LCL = 0\)

单个观测值:数据4

\(UCL=\frac{(m-1)^2}{m}\beta(1-\alpha,\ \frac{p}{2},\ \frac{m-p-1}{2})\)

\(LCL=0\)


\(\alpha\):固定值 \(0.00135\)

\(p\)​:特征数

\(m\):样本数

\(n\):样本大小

\(F\):F 分布 f.ppf(1-p*σ, p, m*n-m-p+1)

\(\beta\):beta 分布 beta.ppf(1-σ, p/2, (m-p-1)/2)



三、描绘点

1. xbar 控制图

数据1:

第一个描绘点:

\(\frac{601.4+601.6+598.0+601.4+599.4}{5} = 600.36\)

第二个描绘点:

\(\frac{600.0+600.2+601.2+598.4+599.0}{5} = 599.76\)

第三个描绘点:

\(\frac{601.2+601.0+600.8+597.6+601.6}{5} = 600.44\)

第四个描绘点:

\(\frac{599.4+601.2+598.4+599.2+598.8}{5} = 599.4\)


2. S 控制图

数据1:

第一个描绘点:

\(\sqrt[2]\frac{[(601.4-600.36)^2+(601.6-600.36)^2+(598.0-600.36)^2+(601.4-600.36)^2+(599.4-600.36)^2]}{4} = 1.596\)

第二个描绘点:

\(\sqrt[2]\frac{[(600-599.76)^2+(600.2-599.76)^2+(601.2-599.76)^2+(598.4-599.76)^2+(599-599.76)^2]}{4} = 1.090\)

第三个描绘点:

\(\sqrt[2]\frac{[(601.2-600.44)^2+(601-600.44)^2+(600.8-600.44)^2+(597.6-600.44)^2+(601.6-600.44)^2]}{4} = 1.615\)

第四个描绘点:

\(\sqrt[2]\frac{[(599.4-599.4)^2+(601.2-599.4)^2+(598.4-599.4)^2+(599.2-599.4)^2+(598.8-599.4)^2]}{4} = 1.077\)


3. R 控制图

数据1:

第一个描绘点:\(601.6-598 = 3.6\)

第二个描绘点:\(601.2-598.4 = 2.8\)

第三个描绘点:\(601.6-597.6 = 4\)


4. 单值控制图

数据1、数据4:就是原始数据


5. 移动极差控制图

数据1:

第二个描绘点:\(601.6-601.4 = 0.2\)

第三个描绘点:\(601.6-598 = 3.6\)

第四个描绘点:\(601.4-598 = 3.4\)

数据4:

第二个描绘点:\(6.05-5.99=0.06\)

第三个描绘点:\(6.11-5.99 = 0.12\)

第四个描绘点:\(6.13-6.11 = 0.02\)


6. MA 控制图

数据1:

第一个描绘点:

\(\frac{601.4+601.6+598.0+601.4+599.4}{5} = 600.36\)

第二个描绘点:

\(\frac{\frac{601.4+601.6+598.0+601.4+599.4}{5}+\frac{600.0+600.2+601.2+598.4+599.0}{5}}{2} = 600.06\)

第三个描绘点:

\(\frac{\frac{601.4+601.6+598.0+601.4+599.4}{5}+\frac{600.0+600.2+601.2+598.4+599.0}{5}+\frac{601.2+601.0+600.8+597.6+601.6}{5}}{mv} = 600.1876\)

第四个描绘点:

\(\frac{\frac{600.0+600.2+601.2+598.4+599.0}{5}+\frac{601.2+601.0+600.8+597.6+601.6}{5}+\frac{599.4+601.2+598.4+599.2+598.8}{5}}{mv} = 600.1876\)

数据4:

第一个描绘点:\(6.05\)

第二个描绘点:\(\frac{6.05+5.99}{2}=6.02\)

第三个描绘点:\(\frac{6.05+5.99+6.11}{mv}=6.05\)

第四个描绘点:\(\frac{5.99+6.11+6.13}{mv}=6.0767\)


7. EWMA 控制图

数据1:

第一个描绘点中的np.mean(datas)是对所有值求平均。

除第一个描绘点外,其他描绘点的计算都与上一个描绘点有关。

第一个描绘点:\(r\times\frac{601.4+601.6+598.0+601.4+599.4}{5}+(1-r)\times(np.mean(datas))=600.130\)

第二个描绘点:\(r\times\frac{600.0+600.2+601.2+598.4+599.0}{5}+(1-r)\times600.13=600.056\)

第三个描绘点:\(r\times\frac{601.2+601.0+600.8+597.6+601.6}{5}+(1-r)\times600.056=600.133\)

第四个描绘点:\(r\times\frac{599.4+601.2+598.4+599.2+598.8}{5}+(1-r)\times600.133=599.986\)

数据4:

第一个描绘点:\(r\times6.05+(1-r)\times(np.mean(datas))=5.9978\)

第二个描绘点:\(r\times5.99+(1-r)\times5.9978=5.9963\)

第三个描绘点:\(r\times6.11+(1-r)\times5.9963=6.0190\)

第四个描绘点:\(r\times6.13+(1-r)\times6.1090=6.0412\)


8. T方控制图

数据1、数据2:

子组ID \(\overline{x}_1\)​(数据1)​ \(\overline{x}_2\)​(数据2)​​​ \(s_{11}\) \(s_{12}\) \(s_{22}\) 统计量 \(T_i^2\)
1 600.36 599.52 2.548 -0.634 0.732 0.281
2 599.76 599.2 1.188 -0.500 0.500 2.283
3 600.44 599.36 2.608 0.422 0.188 0.919
4 599.4 599.56 1.160 0.020 0.388 1.505
5 600.88 599.2 1.152 0.180 0.180 3.734
6 600.32 599.8 2.492 -0.150 0.260 1.238
7 599.8 599.32 1.880 0.440 1.012 1.104
8 598.24 600.12 0.128 0.074 0.552 15.115
9 599.32 599.84 2.192 -0.036 0.108 2.961
10 599.68 599.64 1.252 -0.474 0.328 0.605
11 599.52 599. 1.492 -0.630 0.500 5.907
12 599.32 600.24 2.192 0.484 0.208 8.639
13 599.52 599.08 0.812 -0.282 0.212 4.623
14 600.2 599.2 2.340 -0.240 0.260 1.852
15 601.36 599.68 0.088 0.064 0.132 5.993
16 600.6 599.4 1.060 0.050 0.280 1.185
17 601.12 600.12 0.012 0.002 0.432 9.281
18 599.84 599.6 2.028 0.130 0.100 0.209
19 600.76 599.6 2.128 0.160 0.140 1.662
20 601 599.48 1.300 -0.110 0.092 2.886
均值 \(\overline{\overline{x}}_1\)=600.072 \(\overline{\overline{x}}_2\)=599.548​ \(\overline{\overline{s}}_{11}\)=1.5026 \(\overline{\overline{s}}_{12}\)= -0.0515 \(\overline{\overline{s}}_{22}\)=0.3302

① 样本均值的均值:

\(\overline{\overline{x}}=(\overline{\overline{x}}_1,\ \overline{\overline{x}}_2)'=(600.072,\ 599.548)'\)


② 样本协方差:\(m=20, n=5\)

\(\begin{aligned} & x=[601.4,601.6,598.0,601.4,599.4] \\ & y=[598.0,599.8,600.0,599.8,600.0] \\ & \overline{x}=600.36 \\ & \overline{y}=599.52 \\ \end{aligned}\)

\(\begin{aligned} s_{11} & =\frac{1}{n-1}\sum_{i=1}^n(x_i-\bar{x})^2 \\ & =\frac{1}{n-1}[(601.4-600.36)^2+(601.6-600.36)^2+(598-600.36)^2+(601.4-600.36)^2+(599.4-600.36)^2] \\ & =2.548 \\ \end{aligned}\)

\(\begin{aligned} s_{12} & = \frac{1}{n-1}\sum_{i=1}^n(x_i-\bar{x})(y_i-\bar{y}) \\ & = \frac{1}{n-1}[(601.4-600.36)(598-599.52)+(601.6-600.36)(599.8-599.52)+(598-600.36)(600-599.52)+(601.4-600.36)(599.8-599.52)+(599.4-600.36)(600-599.52)] \\ & = -0.634 \end{aligned}\)

\(s_1= \begin{bmatrix} s_{11} & s_{12}\\ s_{21} & s_{22} \end{bmatrix} = \begin{bmatrix} 2.548 & -0.634\\ -0.634 & 0.732 \end{bmatrix}\)


③ 样本协方差矩阵均值:\(m\)​​ 个矩阵对应位置的均值。

\(\bar{S} = \begin{bmatrix} \bar{s}_{11} & \bar{s}_{12}\\ \bar{s}_{21} & \bar{s}_{22} \end{bmatrix} =\begin{bmatrix} 1.5026 & -0.0515\\ -0.0515 & 0.3302 \end{bmatrix}\)


④ 样本的统计量 \(T_i^2\)

\(\bar{S}^{-1}=\begin{bmatrix} 0.66908978 & 0.10435531\\ 0.10435531 & 3.04474348 \end{bmatrix}\)

\(\begin{aligned} T_i^2 & = n(\bar{x}-\bar{\bar{x}})\bar{S}^{-1}(\bar{x}^T-\bar{\bar{x}})^T \\ & = 5* \begin{bmatrix} 600.36-600.072 & 599.52-599.548 \end{bmatrix}* \bar{S}^{-1}* \begin{bmatrix} 600.36-600.072 \\ 599.52-599.548 \end{bmatrix} \\ & = 5* \begin{bmatrix} 0.288 & -0.028 \end{bmatrix}* \bar{S}^{-1}* \begin{bmatrix} 0.288 \\ -0.028 \end{bmatrix} \\ & = 0.281 \end{aligned}\)


数据5:

子组ID \(x_1\) \(x_2\) \(x_3\) 统计量 \(T_i^2\)
1 1.504 4.075 1.971 3.6011
2 1.685 4.599 2.26 1.3041
3 1.529 4.1 1.994 2.4936
4 1.554 4.19 2.024 1.9272
5 1.604 4.275 2.063 0.9898
6 1.664 4.341 2.12 0.8281
7 1.789 4.981 2.287 2.1348
8 1.723 4.416 2.166 2.2673
9 1.831 5.196 2.285 7.3106
10 1.622 4.353 2.135 0.3211
11 1.683 4.396 2.154 0.7400
12 1.598 4.329 2.18 2.1391
13 1.847 5.168 2.331 4.0995
14 1.793 4.547 2.184 4.9793
15 1.886 5.259 2.389 4.3210
16 1.631 4.338 2.073 1.1237
17 1.543 4.204 2.151 4.0627
18 1.665 4.48 2.282 4.3832
19 1.578 4.349 2.128 1.5162
20 1.533 4.28 2.039 3.6714
21 1.674 4.504 2.192 0.0990
22 1.749 4.371 2.155 5.3129
23 1.83 5.094 2.436 4.4348
24 1.813 4.989 2.428 4.8074
25 1.73 4.396 2.16 3.1322
\(\bar{x}_1\)=1.6823 \(\bar{x}_2\)=4.5292 \(\bar{x}_3\)=2.1835

① 样本均值:
\(\bar{x}=(\bar{x}_1,\ \bar{x}_2,\ \bar{x}_3)'=(1.6823,\ 4.5292,\ 2.1835)'\)

② 样本协方差矩阵:

\(\begin{aligned} & x_1=[1.504,1.685,...,1.813,1.73] \\ & x_2=[4.075,4.599,...,4.989,4.396] \\ & x_2=[1.971,2.26,...,2.428,2.16] \\ & \overline{x}_1=1.6823 \\ & \overline{x}_2=4.5292 \\ & \overline{x}_3=2.1835 \\ \end{aligned}\)

\(\begin{aligned} s_{11} & =\frac{1}{m-1}\sum_{i=1}^m(x_{1i}-\bar{x}_1)^2 \\ & =\frac{1}{m-1}[(1.504-1.6823)^2+(1.685-1.6823)^2+...+(1.813-1.6823)^2+(1.73-1.6823)^2] \\ & =0.0128 \\ \end{aligned}\)

\(\begin{aligned} s_{12} & =\frac{1}{m-1}\sum_{i=1}^m(x_{1i}-\bar{x}_1)(x_{2i}-\bar{x}_2) \\ & =\frac{1}{m-1}[(1.504-1.6823)(4.075-4.5292)+(1.685-1.6823)(4.599-4.5292)+...+(1.73-1.6823)(4.396-4.5292)] \\ & =0.0366 \\ \end{aligned}\)

\(S= \begin{bmatrix} s_{11} & s_{12} & s_{13}\\ s_{21} & s_{22} & s_{23}\\ s_{31} & s_{32} & s_{33}\\ \end{bmatrix}= \begin{bmatrix} 0.0128 & 0.0366 & 0.0123\\ 0.0366 & 0.1298 & 0.0412\\ 0.0123 & 0.0412 & 0.0163\\ \end{bmatrix}\)


③ 样本的统计量 \(T_i^2\)

\(\begin{aligned} T_1^2 & =(x_1-\bar{x})S^{-1}(x_1-\bar{x})^T \\ & = \begin{bmatrix} 1.504-1.6823 & 4.075-4.5292 & 1.971-2.1835 \end{bmatrix} *S^{-1} * \begin{bmatrix} 1.504-1.6823 \\ 4.075-4.5292 \\ 1.971-2.1835 \end{bmatrix} \\ & = \begin{bmatrix} -0.1783 & -0.4542 & -0.2125 \end{bmatrix} *S^{-1} * \begin{bmatrix} -0.1783 \\ -0.4542 \\ -0.2125 \end{bmatrix} \\ &=3.6011 \end{aligned}\)

posted @ 2022-01-27 13:03  做梦当财神  阅读(903)  评论(0编辑  收藏  举报