Scribd
Upload
169 views

Linear Regression Example PDF

Uploaded by

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

Linear Regression Example PDF

Uploaded by

bingo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Linear Regression Example

Ani Katchova

2013 by Ani Katchova. All rights reserved.


Linear Regression Example

Data set: auto


Dependent variable: mpg (miles per gallon)
Independent variable for the simple linear regression: weight
Independent variables for the multiple linear regression: weight, price, and foreign

Estimate a simple and multiple linear regression


Plot regression line
Interpret coefficients and their significance
Estimate the predicted values for the dependent variable
Calculate residuals
35
30
miles per gallon
25
20
15

2 2.5 3 3.5 4 4.5


car weight in 1,000 pounds
Simple linear regression:

mpg (y) Coefficient Standard error t-statistic p-value [95% Confidence Interval]
weight (x) -5.53 0.82 -6.72 0.00 -7.23 -3.83
intercept 38.07 2.61 14.58 0.00 32.68 43.46

Regression line equation: 38.07 5.53


One unit increase in weight (corresponding to a 1,000 pounds) leads to a reduction of 5.53
miles per gallon.
35
30
25
20
15

2 2.5 3 3.5 4 4.5


car weight in 1,000 pounds

y Fitted values
Multiple linear regression
Source SS df MS Number of obs =26
F(3, 22)=15.25
Model 382.1 3 127.4 Prob > F =0
Residual 183.8 22 8.4 R-squared =0.6752
Adj R-squared =0.6309
Total 565.8 25 22.6 Root MSE =2.8902

mpg (y) Coefficient Standard error t-statistic p-value [95% Confidence Interval]
weight -7.12 1.60 -4.44 0.00 -10.45 -3.79
price 0.0002 0.0003 0.85 0.40 -0.0003 0.0008
foreign -2.51 2.06 -1.22 0.24 -6.77 1.76
intercept 42.17 4.26 9.89 0.00 33.32 51.01

From the ANOVA table, the F-statistic is 15.25 and the p-value is less than 0.05. Therefore all
coefficients are jointly significantly different from zero.
The R-square is 0.67 and the adjusted R-squared is 0.63. The regression line has a good fit.
If p-value<0.05 then the coefficient is significantly different from zero. Also if the 95%
confidence interval does not contain zero, the coefficient is significantly different from zero.
Here, only the coefficients on weight and the intercept are significantly different than zero.
If weight is higher by 1 unit (corresponding to 1,000 pounds) then the mpg is 7.12 units lower
in terms of miles per gallon.
Regression models (summary tables)
y=mpg Simple linear Multiple linear
regression regression
Weight -5.5* -7.1*
Price 0.0002
Foreign -2.5
Intercept 38.1* 42.1*
*: Coefficient is significantly different from zero at the 5% significance level.

Interpretation of the coefficients in the simple linear regression model: for cars with 1 unit
higher weight (corresponding to 1,000 pounds more), this is associated with a reduction of 5.5
miles per gallon.
Interpretation of the coefficients in the multiple linear regression model: for cars with 1 unit
higher weight (corresponding to 1,000 pounds more), this is associated with a reduction of
7.1miles per gallon, holding all else constant.
Other coefficients are not significantly different from zero.

You might also like