Multiple Regression
1. The following exercises require the use of a computer and statistical software. Use a 5% significance level.
Perform the Durbin-Watson test at the 5% significance level to determine whether positive first-order autocorrelation exists when d = 1.10, n = 25, and k = 3. You will have two answers for (d). One will be for the subscript (l) and the other for the subscript (u).
Multiple Regressions
Test the following hypotheses with α = .05.
- H0: There is no first-order autocorrelation
- H1: There is positive first-order autocorrelation
- n = 50, k = 2, d = 1.38
Time-Series Analysis and Forecasting 20.19
Plot the data to develop a graph to determine which of the trend models appears to fit better.
|
Period
|
1
|
2
|
3
|
4
|
5
|
|
Time Series
|
55
|
57
|
53
|
49
|
47
|
|
Period
|
6
|
7
|
8
|
9
|
10
|
|
Time Series
|
39
|
41
|
33
|
28
|
20
|
Time-Series Analysis and Forecasting 20.35
- The following trend line and seasonal indexes were computed from 4 weeks of daily observations. Forecast the 7 values for next week. y^ = 120 + 2.3t t = 1, 2, ... , 28
|
Day
|
Seasonal Index
|
|
Sunday
|
1.5
|
|
Monday
|
.4
|
|
Tuesday
|
.5
|
|
Wednesday
|
.6
|
|
Thursday
|
.7
|
|
Friday
|
1.4
|
|
Saturday
|
1.9
|
Add additional columns for (t), the (y) formula shown, and your forecast. Plug in the values indicated and develop your answer.