Inference on the mean of two population , Empirical models Design for

engineering experiments • 3-sigma control limits:

variances known <z-tests import pandas as pd
import matplotlib.pyplot as plt 2k factorial design d
5) M = S D.
.
of sample mean
import numpy as np
from sklearn.linear_model import LinearRegression contrast 601 a ,
b ab
, ,
(7) 27 +, -
&
from sklearn.metrics import mean_squared_error,
mean_absolute_error
eX .
+ k = 2 (a b) ,

x_data = [data1,data2,data3,..]
y_data = [data1,data2,data3,..]
n=num of column
x = np.array(x_data).reshape(-1,1)
y = np.array(y_data).reshape(-1,1)
model = LinearRegression() 22
model.fit(x,y)
contrast
warby In regression model
-sum
=
#y_pred = model.predict(x) >
B11 #model.intercept_ #y-intercept ② Range=MAX(B2:F2)-MIN(B2:F2)
#model.coef_ #slope Sderror effect ⑪ R-bar คือเอาr avgอีกที
var s
95% =
700(7- 2) %
#mean_squared_error(y,y_pred) #MSE >
aug .

mean=AVERAGE(B5:F5)
#model.predict([[ *number* ]]) Grand mean คือเอามีน avgอีกที
0 . 95 = 1 -
2 B14 9 #mean_absolute_error(y,y_pred). #MAE S.d.=STDEV.S(B5:F5)

2 0 05 (9) S barคือเอา s.d. Avgอีกที

⑪
=

data = {'x':[data1,data2,data3,..], 'y':[data1,data2,data3,..]}

.

df = pd.DataFrame(data)
point estimate z(/2) =
NORM S INV11-BM/2) correlation = df['x'].corr(df['y']) ⑪
. .

=
AVG (lower : upper bound print(correlation) ② ทำกราฟอะไรเอาข้อมูลนั้นมาใช้
-

e confidence bound one tail >

-

2x ↓ ⑤
lower bound------ (ztails) --- upper bound eff =
contrast/4 22-18 ·

เอา n=5 ไปหาจากtable

SS =
contrast2/4 22 76 ·

③
ทำกราฟ ครอปแค่ข้อมูล
เลือกline graph >
Inference on the mean of two population ,
variances unknown < +-test) case 2
13 13

Case I df . .
=
n +12 -
2
KG
g
test
statistic
test
statistic .f
d .

df . .
↓

Control Charts for Individual Measurements

um Sp + pooles Wi/1450s
of variance
O
of 2
Sigma(S d
SD
,
14, 3 . .
) 82 (variance
,

9=2

D = 2

case I C = 2 F Avgทั้งแถว >

-
Mr

B12
Seas
& Process Capability
- To B15 • Process Capability Ratio • One-sided Process Capability Ratio - PCRk
t
n -
2/2
Q

> mean square

Paired
t-test <t-tests ~
mean square error

< z-test >

(4/2)
&
② Inference on two population proportions C3 L3
①
d. f .
= n -
1 - sample C3*$L3
propotion

p = total
proportion w
• P Chart (Control Chart for Proportions)
Z
= aver vos all dif
STDEU S a • U Chart (Control Chart for Defects per Unit)
SD =
.

SUM(แถวส้ม)
w w Contrast /2*2^(3-1)
Inference the ratio of variances of two < F-test > Contrast^2 /2*2^3
on ② ⑪ U bar is avg number of defect per unit**
normal populations ⑤ • EWMA Control Chart
confidence interval on
dif. in true propotions p . -P2
⑤ S
~ confidence interval
on the ratio
02/02
B12
Tail He 2,2 re
2 tails
fractional 2
designO 8
ex
replication
=
.

(P)
*
[f() alpha) -

use in Zo G
=F.INV.RT(1-alpha,n(1)-1,n(2)-1)
B76
disten
contrast
* S Effect =

B19
=1-F.DIST(E54,1,$C$59,TRUE)
2nk15gy (independents -
1
SSE
* dentalition ! & 159 no now rosvovado wan total
*