Lesson8 PDF
Lesson8 PDF
Abdelfatah Mustafa
Page 1 of 81
Multiple Regression
MultipleRegression
Multiple
Page 2 of 81
1-8 Introduction
.
.
) (
)( .
Regression Equation
Page 3 of 81
Exact Approximate
)( .
.
Linear
Non
Non-Linear
Linear
.
Page 4 of 81
SPSS
2-8
Models
( .
Page 5 of 81
.Categorical
.
Discriminant analysis
y
Regression Models
Page 6 of 81
)(
.1
.2
:
)( Standard or Simultaneous Regression
Page 7 of 81
.
Hierachical Regression
g
)(
.
St
Stepwise
i Regression
R
i
)(
) (
.
Criteria
.
[email protected]
Page 8 of 81
forward backward
.
:
F-Test
.
F-test
F test
FIN
PIN .
Page 9 of 81
:
F F
F-Test
Test
FOUT POUT .
:
.
.
:
[email protected]
Page 10 of 81
.1
to
cases
or
Ratio
independent variables
.
.
.
.
.2
Outliers
Page 11 of 81
Mahalanobis
.
SPSS
.
.3
Multicollinearity Singularity
.
.
Page 12 of 81
Correlation Matrix
Tolerances
.4
Normality
Linearity
Homoscedasticity Independence of
Residuals
Page 13 of 81
)(i
)) ((ii
.
.
) (iii
.SPSS
Page 14 of 81
Years
Temper
ature
Sales
21
15
18
15
22
21
24
28
25
30
15
25
35
26
40
34
35
25
30
38
45
40
50
41
60
39
45
37
60
40
50
Page 15 of 81
.
Regression.sav
sales, temp, years
.SPSS
SPSS
:
.1
.2
2
Stepwise Regression
.3
Hierachical Regression
.4
Page 16 of 81
3-8
MultipleLinearRegression
Page 17 of 81
m X1,
.m =1
X2, , Xm 1
n
Y j = 0 + 1 X 1 j + 2 X 2 j + L + m 1 X mj + j , j = 1, 2, ..., n
Yj m
i= 1,
1 2 , m,
m j=1,
j=1 2,
Xij 2 , n
[email protected]
Page 18 of 81
Parameters j =j
1, 2, , m
j
:
.1
.2
.3
.4
4
Page 19 of 81
Least
Square
.
.
Yj = 0 + 1 X 1 j + 2 X 2 j + L + m 1 X mj , j = 1, 2, ..., n
j , j = 1, 2, ..., m 1
Page 20 of 81
.(
Page 21 of 81
SPSS
regression.sav
:
.1
Analyze Regression
.2
Linear
.3
Linear Regression
Page 22 of 81
Dependent Sales
.4
Independent temp,
temp years
.5
5
Method
.6
Enter
Page 23 of 81
Linear Statistics
Regression: Statistics
.7
Page 24 of 81
.8
Continue
Linear
.9
Plot .10
Regression: Plots
Page 25 of 81
Continue
.13
Page 26 of 81
Linear Save
.14
Regression: Save
Distances
.15
Mahalanobis
Continue
.16
.
Ok .17
.
Page 27 of 81
:
Descriptive Statistics
.
Page 28 of 81
:
Correlations
.
Variables Entered/Removed
year, temp
Page 29 of 81
Model Summary
.1
1
R=0.912
R=0 912
.2
) (
%80 .
.3
Adjusted R Square
.
.4
Page 30 of 81
:
ANOVA
: ) (
(
: )
0.05
.
[email protected]
Page 31 of 81
:
Coefficients
.1
1
Unstandarized Coefficients
Page 32 of 81
.2
Sig.
constant years
constant,
sig 0.05
0.05
sig. 0 05
i =0.016
0 016
.3
.
.
Page 33 of 81
:
Coefficient Correlations
.
:
Residuals Statistics
Page 34 of 81
Page 35 of 81
:
(
)
Page 36 of 81
Page 37 of 81
mah_1
Chi-Square Mahalanobis
n-1 = 2
[email protected]
Page 38 of 81
Page 39 of 81
4-8
Stepwise Regression
StepwiseRegression
Page 40 of 81
.
.
.
)
(.....
[email protected]
Page 41 of 81
) (
.
.
:
.1
.2
.3
Page 42 of 81
.4
.5
) Coefficient of Determination (
F
.6
F
) (.
:
.1
.2
Page 43 of 81
SPSS
: regression.sav
[email protected]
Page 44 of 81
Analyze
.1
Regression
Linear
Linear
.2
.3
regression
Sales
.4
Dependent
temp, year
Independent
Method
.5
Stepwise
[email protected]
Page 45 of 81
Linear Statistics
.6
Regression: Statistics
Model fit,
fit Estimates,
Estimates
.7
7
diagnostics Casewise
Continue
.8
Plot
Pl t
.9
9
Linear
Regression: Plots
[email protected]
Page 46 of 81
* ZPRED
Y:
* ZRESID
.11
X:
Histogram Standardized Residual Plots .12
Histogram,
12
Normal probability plot
Continue
C ti
.13
13
Linear
Save
.14
14
Regression: Save
[email protected]
Page 47 of 81
Distances
.15
Mahalanobis
Continue
.16
16
Ok .17
.
Page 48 of 81
Variables Entered/removed
temp
.
Page 49 of 81
:
Model Summary
) (
80% .
:
ANOVA
0.05
Sig.
g = 0.00
Page 50 of 81
:
Coefficients
S l = 11.271
Sales
2 1 + 1.600 temp
Page 51 of 81
Excluded Variables
:
.
Residuals Statistics
Page 52 of 81
Page 53 of 81
:
)(
.
[email protected]
Page 54 of 81
Page 55 of 81
5-8
Hierachical Regression
Page 56 of 81
regression.sav
:
.1
Analyze
Regression
.2
.3
Linear
Regression
Linear
Page 57 of 81
Sales
.4
D
Dependent
d t
temp
Independent(s)
Block 2 of 2 previous
Next
.5
5
years
.6
Independent
.7
Previous
Next
Method
.8
8
Enter
[email protected]
Page 58 of 81
Linear Statistics
.9
Regression: Statistics
Model fit, Estimates,
diagnostics, R
.10
Casewise
Squared Change
Continue
.11
Linear
Page 59 of 81
Plot .12
R
Regression:
i
Pl t
Plots
Continue .15
[email protected]
Page 60 of 81
Linear Save
.16
Regression: Save
Distances
.17
17
Mahalanobis
Continue
.18
.
Ok
.19
Page 61 of 81
:
Variables Entered/Removed
Page 62 of 81
:
R
M d l Summary
Model
S
:
ANOVA
.
[email protected]
Page 63 of 81
:
)(
Coefficients
. :
.1
1
temp .
.2
Page 64 of 81
Excluded Variables
Page 65 of 81
:
. Residual Statistics
Page 66 of 81
:
.
Page 67 of 81
:
)(
.
Page 68 of 81
6-8
Curve
CurveLinearRegression
Linear Regression
Page 69 of 81
....
....
SPSS
:
.1
1
C bi
Cubic
Quadratic
Q d i
Linear
Li
.2
Exponential
.3
Logarithmic
.4
Inverse Compound
.5
Logistic Power
Page 70 of 81
Regression.sav
SPSS
:
.1
.2
Analyze Regression
Curve
Estimation
.3
Curve Estimation
Sales Dependent years
Independent Variable
Page 71 of 81
Time
.4
Exponential
.5
I l d constant
Include
t t in
i Equation,
E
ti
Pl t Models
Plot
M d l
.6
Page 72 of 81
Display ANOVA
.7
Save
.8
Page 73 of 81
.9
Continue
C ti
.10
10
Page 74 of 81
: Ok .10
CurveFit
Page 75 of 81
:
Model Description
.
.()
Page 76 of 81
:
Variable Processing Summary
)
.(
Page 77 of 81
Sales: Exponential
Sales:Exponential
Model Summary
.
Page 78 of 81
ANOVA
( .
:
Coefficients
.
0.23419 years
sales =17.25 e
Page 79 of 81
:
.
regression.sav
[email protected]
Page 80 of 81
:
Fit_1
Fit_
1: Fit for SALES with YEARS from CURVEFIT, Mod_1 EXPONENTIAL
ERR_1
ERR_
1: Error for SALES with YEARS from CURVEFIT,
Mod_1
EXPONENTIAL
[email protected]
Page 81 of 81
Mod_1
EXPONENTIAL
UCL
UCL_1
UCL_
1: 95% UCL for SALES with YEARS from CURVEFIT,
CURVEFIT
Mod 1
Mod_1
EXPONENTIAL
.