In: Statistics and Probability
MWH | CUS | AHEC | t | D1 | D2 | |
Jan-10 | 221,870 | 329,075 | 674.2 | 1 | 1 | 0 |
Feb-10 | 175,726 | 329,754 | 532.9 | 2 | 0 | 0 |
Mar-10 | 158,684 | 330,374 | 480.3 | 3 | 0 | 0 |
Apr-10 | 179,014 | 330,751 | 541.2 | 4 | 0 | 0 |
May-10 | 158,494 | 331,128 | 478.6 | 5 | 0 | 0 |
Jun-10 | 245,588 | 331,050 | 741.8 | 6 | 0 | 1 |
Jul-10 | 285,503 | 332,673 | 858.2 | 7 | 0 | 1 |
Aug-10 | 283,385 | 332,918 | 851.2 | 8 | 0 | 1 |
Sep-10 | 250,406 | 333,162 | 751.6 | 9 | 0 | 0 |
Oct-10 | 211,725 | 333,284 | 635.3 | 10 | 0 | 0 |
Nov-10 | 156,017 | 333,518 | 467.8 | 11 | 0 | 0 |
Dec-10 | 182,422 | 334,508 | 545.3 | 12 | 1 | 0 |
Jan-11 | 213,135 | 335,497 | 635.3 | 13 | 1 | 0 |
Feb-11 | 194,887 | 335,813 | 580.3 | 14 | 0 | 0 |
Mar-11 | 133,260 | 336,405 | 396.1 | 15 | 0 | 0 |
Apr-11 | 154,670 | 336,637 | 459.5 | 16 | 0 | 0 |
May-11 | 171,713 | 336,869 | 509.7 | 17 | 0 | 0 |
Jun-11 | 310,874 | 336,918 | 922.7 | 18 | 0 | 1 |
Jul-11 | 318,138 | 337,519 | 942.6 | 19 | 0 | 1 |
Aug-11 | 304,805 | 338,063 | 901.6 | 20 | 0 | 1 |
Sep-11 | 276,765 | 338,606 | 817.4 | 21 | 0 | 0 |
Oct-11 | 207,074 | 338,987 | 610.9 | 22 | 0 | 0 |
Nov-11 | 159,807 | 339,361 | 470.9 | 23 | 0 | 0 |
Dec-11 | 188,262 | 340,009 | 553.7 | 24 | 1 | 0 |
Jan-12 | 227,781 | 340,656 | 668.7 | 25 | 1 | 0 |
Feb-12 | 159,576 | 341,266 | 467.6 | 26 | 0 | 0 |
Mar-12 | 168,212 | 341,912 | 492.0 | 27 | 0 | 0 |
Apr-12 | 169,883 | 342,461 | 496.1 | 28 | 0 | 0 |
May-12 | 217,582 | 342,850 | 634.6 | 29 | 0 | 0 |
Jun-12 | 285,227 | 343,216 | 831.0 | 30 | 0 | 1 |
Jul-12 | 320,056 | 343,877 | 930.7 | 31 | 0 | 1 |
Aug-12 | 313,702 | 344,271 | 911.2 | 32 | 0 | 1 |
Sep-12 | 251,051 | 344,418 | 728.9 | 33 | 0 | 0 |
Oct-12 | 177,654 | 345,105 | 514.8 | 34 | 0 | 0 |
Nov-12 | 163,067 | 345,305 | 472.2 | 35 | 0 | 0 |
Dec-12 | 194,556 | 345,567 | 563.0 | 36 | 1 | 0 |
Jan-13 | 247,126 | 345,667 | 714.9 | 37 | 1 | 0 |
Feb-13 | 184,078 | 346,239 | 531.7 | 38 | 0 | 0 |
Mar-13 | 167,302 | 346,554 | 482.8 | 39 | 0 | 0 |
Apr-13 | 144,089 | 346,943 | 415.3 | 40 | 0 | 0 |
May-13 | 206,792 | 347,271 | 595.5 | 41 | 0 | 0 |
Jun-13 | 308,944 | 347,866 | 888.1 | 42 | 0 | 1 |
Jul-13 | 316,351 | 347,982 | 909.1 | 43 | 0 | 1 |
Aug-13 | 308,119 | 348,610 | 883.9 | 44 | 0 | 1 |
Sep-13 | 255,635 | 349,077 | 732.3 | 45 | 0 | 0 |
Oct-13 | 168,472 | 349,352 | 482.2 | 46 | 0 | 0 |
Nov-13 | 166,379 | 349,497 | 476.1 | 47 | 0 | 0 |
Dec-13 | 205,975 | 349,629 | 589.1 | 48 | 1 | 0 |
Jan-14 | 230,231 | 349,921 | 658.0 | 49 | 1 | 0 |
Feb-14 | 164,773 | 350,111 | 470.6 | 50 | 0 | 0 |
Mar-14 | 148,026 | 350,967 | 421.8 | 51 | 0 | 0 |
Apr-14 | 148,877 | 351,667 | 423.3 | 52 | 0 | 0 |
May-14 | 202,760 | 352,097 | 575.9 | 53 | 0 | 0 |
Jun-14 | 298,366 | 352,340 | 846.8 | 54 | 0 | 1 |
Jul-14 | 356,241 | 352,727 | 1010.0 | 55 | 0 | 1 |
Aug-14 | 300,386 | 352,858 | 851.3 | 56 | 0 | 1 |
Sep-14 | 237,898 | 353,640 | 672.7 | 57 | 0 | 0 |
Oct-14 | 193,688 | 353,434 | 548.0 | 58 | 0 | 0 |
Nov-14 | 151,736 | 353,649 | 429.1 | 59 | 0 | 0 |
Dec-14 | 207,533 | 353,885 | 586.4 | 60 | 1 | 0 |
Jan-15 | 236,812 | 354,068 | 668.8 | 61 | 1 | 0 |
Feb-15 | 159,097 | 354,644 | 448.6 | 62 | 0 | 0 |
Mar-15 | 165,744 | 355,563 | 466.1 | 63 | 0 | 0 |
Apr-15 | 150,227 | 355,971 | 422.0 | 64 | 0 | 0 |
May-15 | 194,866 | 356,579 | 546.5 | 65 | 0 | 0 |
Jun-15 | 295,847 | 356,932 | 828.9 | 66 | 0 | 1 |
Jul-15 | 366,289 | 357,572 | 1024.4 | 67 | 0 | 1 |
Aug-15 | 342,595 | 357,750 | 957.6 | 68 | 0 | 1 |
Sep-15 | 292,113 | 358,421 | 815.0 | 69 | 0 | 0 |
Oct-15 | 208,374 | 358,514 | 581.2 | 70 | 0 | 0 |
Nov-15 | 138,642 | 358,803 | 386.4 | 71 | 0 | 0 |
Dec-15 | 220,531 | 358,819 | 614.6 | 72 | 1 |
0 |
MWH = | total MWH consumed by all Texas and New Mexico households served by EPE in month t | (MWH = megawatt hours) | ||||||||
CUS = | total number of residential customers in month t | |||||||||
AHEC = | kWh average household electricity consumtion in month t = (MWH/CUS)*1000 | (kWh = kilowatt hours) | ||||||||
D1 = | 1 if month t pertains to December or January; 0 otherwise | |||||||||
D2 = | 1 if month t pertains to June, July or August; 0 otherwise |
(i) Estimate the following time-series model
?????=?1+?2∙?+ ?3∙?1?+?4∙?2?
?????= average household electricity consumption (kWh) in month t in El Paso Electric’s combined Texas and New Mexico service territory.
?1?= 1 if month t pertains to December or January (winter heating peak); 0 otherwise
?2?= 1 if month t pertains to June, July or August (summer cooling peak); 0 otherwise
(ii) With a two-tailed test (and a = 0.05), test the hypothesis that in the population ?2=0 (iii) With a two-tailed test (and a = 0.05), test the hypothesis that in the population ?3=0
(iv) With a two-tailed test (and a = 0.05), test the hypothesis that in the population ?4=0
(v) Based on your results from part (ii), does there appear to be either a significant upward or downward time trend in average household electricity consumption in El Paso Electric’s combined Texas and New Mexico service territory?
(vi) Interpret the R2 statistic.
(vii) Provide a forecast for AHEC for each of the following months: (1) January 2016 (t = 73); (2) July 2016 (t = 79); and (3) October 2016 (t = 82).
(viii) Given actual values in the source references found in “EPE Sales Cus” (click on cells B-82 and B-83), calculate the forecast errors associate with each of the three forecasts in part (vii) and note if they were underestimates or overestimates
(ix) As it turned out, July 2016 was 35% hotter than normal (based on “cooling degree days”). Why might this explain the July 2016 forecast error from part (viii)?
1 - 5 have been solved.
Input the data in excel as shown. Using the regression tool from
data analysis tab, the regression output is generated.
Inputs for the regression tab are shown below.
The regression output is as shown below.
The regression equation is given as
y = 531.7904 +0.08307 t+87.844 D1+359.085 D2
(ii) With a two-tailed test (and a = 0.05), test the hypothesis
that in the population ?2=0
From the regression table we have
Using the t-table, with df = 72-4 = 68, we get pvalue = 0.8777.
Since the pvalue is greater than 0.05, we fail to reject the null hypothesis and conclude that beta2 is not significant in predicting y.
(iii) With a two-tailed test (and a = 0.05), test the hypothesis that in the population ?3=0
Using the t-table, with df = 72-4 = 68, we get pvalue = 0.006117716
Since the pvalue is less than 0.05, we reject the null hypothesis
and conclude that beta3 is significant in predicting y.
(iv) With a two-tailed test (and a = 0.05), test the hypothesis that in the population ?4=0
Using the t-table, with df = 72-4 = 68, we get pvalue = 0.000000008
Since the pvalue is less than 0.05, we reject the null hypothesis
and conclude that beta4 is significant in predicting y.
(v) Based on your results from part (ii), does there appear to
be either a significant upward or downward time trend in average
household electricity consumption in El Paso Electric’s combined
Texas and New Mexico service territory?
Slope of the regression line is positive. There is a signficant
increase int he electricity in a seasonal pattern.
(vi) Interpret the R2 statistic.
The R2 = 0.7270, that is 72.70%
In simple words, it means 72% of the variation in the data is being
explained by the independent variables.