The manager of an amusement park would like to be able to predict daily attendance to develop more accurate plans about how much food to order and how many ride operators to hire. After some consideration, he decided that the following three factors are critical:
He then took a random sample of 40 days. For each day, he recorded the attendance, the previous day’s attendance, day of the week, and weather forecast. An example of the first few lines of Data and the regression output are below:
Regression Statistics |
||||||
Multiple R |
0.836766353 |
|||||
R Square |
0.700177929 |
|||||
Adjusted R Square |
0.665912549 |
|||||
Standard Error |
810.7745532 |
|||||
Observations |
40 |
|||||
ANOVA |
||||||
df |
SS |
MS |
F |
Significance F |
||
Regression |
4 |
53729535 |
13432384 |
20.43398 |
9.28E-09 |
|
Residual |
35 |
23007438 |
657355.4 |
|||
Total |
39 |
76736973 |
||||
Coefficients |
Standard Error |
t Stat |
P-value |
Lower 95% |
Upper 95% |
|
Intercept |
3490.466604 |
469.1554 |
7.439894 |
1.04E-08 |
2538.031 |
4442.903 |
YestAtt |
0.368547078 |
0.077895 |
4.731349 |
3.6E-05 |
0.210412 |
0.526682 |
I1 |
1623.095785 |
492.5497 |
3.295294 |
0.002258 |
623.1668 |
2623.025 |
I2 |
733.4646317 |
394.3718 |
1.85983 |
0.071331 |
-67.1527 |
1534.082 |
I3 |
765.5429068 |
484.6621 |
-1.57954 |
0.123209 |
-1749.46 |
218.3734 |
If the manager is looking for a way to help predict attendance, Is this a good model to use? How would you suggest making this model better? please give a detailed answer with values.
To get the validity of the model , the F test is used
Here the F statistic is significant which implies that our model fit the data well than the intercept only model .
To check whether weather is a factor in determining the attendance, t test is used
Here p value(.123)for weather is greater than .05, we conclude that the variable weather does not affect the attendance significantly.
This model has R2 value(.7002) , which implies it is good model to use but removing the insignificant variables weather and type of day , we can get even a better model to predict the attendance well
Get Answers For Free
Most questions answered within 1 hours.