Download as xlsx, pdf, or txt
Download as xlsx, pdf, or txt
You are on page 1of 15

Notes on Using the Model

We have two different models that predict mileage based on different input parameters. Both the
models are regression-based models with different levels of accuracy in predicting mileage. Both
have input variables based on how car mileage is estimated. Play around with input variables and
see how it affects the car mileage. Ultimately you will use these models to answer questions
provided in the Graded Activity page for Week 1.

Model 1: Uses information about number of Cylinders, Engine Displacement, Drive, Transmission
Type, Vehicle Class and Internal Volume of the car as parameters to estimate the expected vehicle
mileage. The model is around 77% accurate

Model 2: Uses the CO2 emissions from the car to predicts expected vehicle mileage. The model
results are accurate 92 times out of 100

Note: Please do not update the equation for the Output variable for any of the models.

Regression Result Model 1 This sheet contains the results of the regression analysis of the model b
Regression Result Model 2 This sheet contains the results of the regression analysis of the model b

Car Mileage Data This sheet contains the test data on which the model is based. This data
arameters. Both the
cting mileage. Both
input variables and
wer questions

Drive, Transmission
he expected vehicle

leage. The model

n analysis of the model based on the input variables for the first m
n analysis of the model based on the input variables for the secon

model is based. This data as mentioned above is collected from the US Dept of Energy web
Click on the shaded orange cells below to change model parameters and observe how the predicted mileage is affected

Input Variables
Cylinders 6 units
Engine Displacement 1.4 liters Acceptable displacement is a decimal with a range betwee
Drive 4 wheel drive
Car Internal Volume 320 cubic feet Acceptable volume is an integer with a range between 0 a
Transmission Manual
Vehicle Class Van

Expected Output
Predicted Mileage 24.12 mpg

24.25
25.63
24.12
edicted mileage is affected.

ecimal with a range between 0.2 and 12.

r with a range between 0 and 800.


Double-click on the shaded orange cell below to change model parameters and observe how the predicted milea

Input Variables
Co2 Emissions 1500 grams/mile Acceptable input is an integer with a range betw

Expected Output
Predicted Mileage 0.00 mpg
bserve how the predicted milea

ble input is an integer with a range between 0 and 1600


The model predicts the mileage of the vehicle basd on the number of cylinders, engine
This sheet contains results of the regression analysis for our car mileage model based o
being used. If you are not familiar with regression analysis, then you may be interested
not required for this activity.
Regression Statistics
Multiple R 0.777
R Square 0.604
Adjusted R Square 0.604
Standard Error 3.222
Observations 37731

ANOVA
df SS MS F Significance F
Regression 6 596466.284 99411.047 9577.275 0
Residual 37724 391570.917 10.380
Total 37730 988037.201

Coefficients Standard Error t Stat P-value Lower 95% Upper 95%


Intercept 30.214 0.088 344.451 0 30.043 30.386
Cylinders -0.670 0.023 -29.523 0.000 -0.715 -0.626
Engine Displacement -1.707 0.030 -56.484 0 -1.766 -1.648
Drive -0.448 0.020 -22.318 0.000 -0.487 -0.409
Car Internal Volume 0.010 0.000 34.470 0.000 0.010 0.011
Transmission 0.068 0.037 1.850 0.064 -0.004 0.140
Vehicle Class -0.200 0.010 -20.641 0.000 -0.220 -0.181
mber of cylinders, engine displacement, Drive, Transmission Type , Total Internal Volume, Vehicle Class
r mileage model based on engine parameters. This sheet might help to figure out the accuracy level of the model
n you may be interested in reading about R-squared and regression statistics. A detailed knowledge of regression is

Lower 95.0% Upper 95.0%


30.043 30.386
-0.715 -0.626
-1.766 -1.648
-0.487 -0.409
0.010 0.011
-0.004 0.140
-0.220 -0.181
vel of the model
dge of regression is
The model predicts the mileage of the vehicle basd on the CO2 Emissions from the vehicle
This sheet contains results of the regression analysis for our car mileage model based on engine param
are not familiar with regression analysis, then you may be interested in reading about R-squared and re

Regression Statistics
Multiple R 0.924
R Square 0.854
Adjusted R Squar 0.854
Standard Error 1.955
Observations 37731

ANOVA
df SS MS F Significance F
Regression 1 843808.8537 843808.8537 220733.7522 0
Residual 37729 144228.3471 3.823
Total 37730 988037.2008

Coefficients Standard Error t Stat P-value Lower 95% Upper 95%


Intercept 38.767 0.041 939.752 0 38.686 38.848
CO2 Emissions -0.040 0.000 -469.823 0 -0.040 -0.039
m the vehicle
el based on engine parameters. This sheet might help to figure out the accuracy level of the model being used. If you
ng about R-squared and regression statistics. A detailed knowledge of regression is not required for this activity.

Lower 95.0% Upper 95.0%


38.686 38.848
-0.040 -0.039
el being used. If you
or this activity.
comb08 co2Tailpip cylinders displ drive TotalVolu trany
7 1269.5714 12 5.2 2 53 0
7 1269.5714 12 5.2 2 53 0
7 1269.5714 12 5.2 2 53 0
7 1269.5714 12 5.2 2 53 0
7 1269.5714 12 5.2 2 0 0
VClass year make model
0 1986 LamborghinCountach
0 1987 LamborghinCountach
0 1988 LamborghinCountach
0 1989 LamborghinCountach
0 1990 LamborghinCountach

You might also like