Scribd
Upload
111 views

Multiple Regression Analysis

The document provides instructions and a sample dataset to calculate the multiple regression line, standard error of estimate, and coefficient of determination for a multiple regression model with three variables. It shows the calculations for the partial regression coefficients, intercept, standard error of estimate, and coefficient of determination. The key steps are to calculate the sums of squares and cross products, use the normal equations to determine the regression coefficients, then use those coefficients to calculate the standard error and coefficient of determination.

Uploaded by

Aleena
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
111 views

Multiple Regression Analysis

The document provides instructions and a sample dataset to calculate the multiple regression line, standard error of estimate, and coefficient of determination for a multiple regression model with three variables. It shows the calculations for the partial regression coefficients, intercept, standard error of estimate, and coefficient of determination. The key steps are to calculate the sums of squares and cross products, use the normal equations to determine the regression coefficients, then use those coefficients to calculate the standard error and coefficient of determination.

Uploaded by

Aleena
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Practice for multiple regression analysis GCUF-BBA

Suppose three variables:-

a. Find the multiple regression line X3 on X2 and X1 using formulae.


b. Compute the standard error of estimate.
c. Calculate coefficient of multiple determination.

1st 2nd 3rd


variable variable variable

2 1 12

3 2 10

0 3 9

2 4 13

4 5 16

Solution:

X1 X2 X3 X1X2 X1X3 X2X3 𝑿𝟏 𝟐 𝑿𝟐 𝟐 𝑿𝟑 𝟐


2 1 12 2*1=2 2*12=24 12 𝟐𝟐 =4 𝟏𝟐 =1 𝟏𝟐𝟐 =144
3 2 10 6 30 20 9 4 100
0 3 9 0 0 27 0 9 81
2 4 13 8 26 52 4 16 169
4 5 16 20 64 80 16 25 256
TOTAL 11 15 60 36 144 191 33 55 750

Using formulae:

(∑ 𝑋1 )2
∑ 𝑥12 = ∑ 𝑋12 −
𝑛 # of observations given in data set,
in this present case is “n=5”
∑ 𝑥12 = ∑ 𝑋12 −

(∑ 𝑋1 )2 (11)2
∑ 𝑥12 = ∑ 𝑋12 − =33 − =8.8
𝑛 5

(∑ 𝑋2 )2 (15)2 (60)2
∑ 𝑥22 = ∑ 𝑋22 − =55 − =10 , ∑ 𝑥32 = 750 − = 30
𝑛 5 5
∑ 𝑋1 ∑ 𝑋2 (11)(15)
∑ 𝑥1 𝑥2 = ∑ 𝑋1 𝑋2 − = 36 − =3
𝑛 5

∑ 𝑋1 ∑ 𝑋3 (11)(60)
∑ 𝑥1 𝑥3 = ∑ 𝑋1 𝑋3 − = 144 − =12 , ∑ 𝑥2 𝑥3 = 11
𝑛 5

(∑ 𝑥1 𝑥3 )(∑ 𝑥22 )−(∑ 𝑥1 𝑥2 )(∑ 𝑥2 𝑥3 ) 12∗10−11∗3 87


𝑏1 = ∑ 𝑥12 ∑ 𝑥22 −(∑ 𝑥1 𝑥2 )2
= = 79=1.10 Partial regression
8.8∗10−32
coefficients
(∑ 𝑥2 𝑥3 )(∑ 𝑥12 )−(∑ 𝑥1 𝑥2 )(∑ 𝑥1 𝑥3 ) 11∗8.8−3∗12 60.8
𝑏2 = ∑ 𝑥12 ∑ 𝑥22 −(∑ 𝑥1 𝑥2 )2
= = =0.77
8.8∗10−32 79

Intercept of multiple
𝑎 = 𝑋̅3 − 𝑏1 𝑋̅1 − 𝑏2 𝑋̅2 = 12 − (1.10)(2.2) − (0.77)(3)=7.27 regression

Hence required multiple regression line X3 on X2 and X1 is:

𝑥̂3 = 𝑎 + 𝑏1 𝑥1 + 𝑏2 𝑥2 = 7.27 + 1.10𝑥1 + 0.77𝑥2

Standard error of estimate:

∑ 𝑋32 − 𝑎 ∑ 𝑋3 − 𝑏1 ∑ 𝑋1 𝑋3 − 𝑏2 ∑ 𝑋2 𝑋3
𝑠3.12 = √
𝑛−3

750 − 7.27(60) − 1.10(144) − 0.77(191)


𝑠3.12 = √
5−3

8.33
𝑠3.12 = √ = √4.165 = 2.04
2
Coefficient of determination:

∑(𝑋̂ −𝑋̂ )2
𝑅3.21 = ∑(𝑋3 −𝑋̿3)2
3 3

(∑ 𝑋3 )2
∑(𝑋̂3 − 𝑋̂3 )2 = 𝑎 ∑ 𝑋3 + 𝑏1 ∑ 𝑋1 𝑋3 + 𝑏2 ∑ 𝑋2 𝑋3 −
𝑛
2
60
= 7.27 ∗ 60 + 1.10 ∗ 144 + 0.77 ∗ 191 − = 21.67
5

(∑ 𝑋3 ) 2
∑(𝑋3 − 𝑋̿3 )2 = ∑ 𝑋32 − = 30
𝑛

21.67
𝑅3.12 = = 0.7223
30

Carryout the necessary computation to obtain least square estimate of the parameters (partial
regression coefficient and intercept) in multiple regression. Also compute goodness of fit and
standard error of estimate. Consider 1st and 3rd variable as dependent variable and provide all
required computation.

1st 12 10 07 13 20 11 09 08
2nd 2 2 3 4 4 5 1 6
3rd 1 1 0 0 3 10 6 4

You might also like