In: Statistics and Probability
According to the Energy Information Administration, the United States consumes about 21 million barrels (882 million gallons) of petroleum each day. About half of this consumption is in the form of gasoline. Table 18.2 shows a portion of weekly U.S. finished motor gasoline production, measured in thousands of barrels per day.
1. Construct a 3-period moving average series for the data.
2.) Plot production and its corresponding 3-period moving average, and comment on any differences.
3.) Using the 3-period moving average series, forecast gasoline production on May 29, 2009 (week 22).
4.) Calculate the mean square error, MSE, and the mean absolute deviation,
Date | Week | Production |
January 2, 2009 | 1 | 9115 |
January 9, 2009 | 2 | 8813 |
January 16, 2009 | 3 | 8729 |
January 23, 2009 | 4 | 8660 |
January 30, 2009 | 5 | 8679 |
February 6, 2009 | 6 | 8492 |
February 13, 2009 | 7 | 8765 |
February 20, 2009 | 8 | 8937 |
February 27, 2009 | 9 | 9003 |
March 6, 2009 | 10 | 8539 |
March 13, 2009 | 11 | 8868 |
March 20, 2009 | 12 | 8723 |
March 27, 2009 | 13 | 8733 |
April 3, 2009 | 14 | 8966 |
April 10, 2009 | 15 | 8913 |
April 17, 2009 | 16 | 9088 |
April 24, 2009 | 17 | 8790 |
May 1, 2009 | 18 | 8918 |
May 8, 2009 | 19 | 8710 |
May 15, 2009 | 20 | 8735 |
May 22, 2009 | 21 | 9378 |
1. Construct a 3-period moving average series for the data.
Ans:
Week | Production | 3-period moving average |
1 | 9115 | |
2 | 8813 | 8885.667 |
3 | 8729 | 8734 |
4 | 8660 | 8689.333 |
5 | 8679 | 8610.333 |
6 | 8492 | 8645.333 |
7 | 8765 | 8731.333 |
8 | 8937 | 8901.667 |
9 | 9003 | 8826.333 |
10 | 8539 | 8803.333 |
11 | 8868 | 8710 |
12 | 8723 | 8774.667 |
13 | 8733 | 8807.333 |
14 | 8966 | 8870.667 |
15 | 8913 | 8989 |
16 | 9088 | 8930.333 |
17 | 8790 | 8932 |
18 | 8918 | 8806 |
19 | 8710 | 8787.667 |
20 | 8735 | 8941 |
21 | 9378 |
2.) Plot production and its corresponding 3-period moving average, and comment on any differences.
Comment The corresponding 3-period moving average has slightly smoother than the actual data plot.
3.) Using the 3-period moving average series, forecast gasoline production on May 29, 2009 (week 22).
Ans: Using the 3-period moving average series, forecast gasoline production on May 29, 2009 is 8941.
4.) Calculate the mean square error, MSE, and the mean absolute deviation,
Ans:
MSE=MSD = 24172.7
mean absolute deviation=MAD = 117.5
s