The data shown below represent the monthly sales of a product over 5 years. 1 2
ID: 1133372 • Letter: T
Question
The data shown below represent the monthly sales of a product over 5 years.
1
2
3
4
5
Jan
1030
896
951
991
1042
Feb
1032
793
861
950
997
Mar
1126
885
938
1024
1076
Apr
1285
1055
1109
1182
1198
May
1468
1204
1274
1349
1330
Jun
1637
1326
1422
1473
1407
Jul
1611
1303
1486
1540
1465
Aug
1608
1436
1555
1599
1516
Sep
1528
1473
1604
1591
1508
Oct
1420
1453
1600
1546
1431
Nov
1119
1170
1403
1316
1271
Dec
1013
1023
1209
1151
1083
You have developed a new forecasting model for this product, and wish to test your model against one of the simple forecasting methods.
(a) Which simple forecasting method will you use as a comparison?
(b) Provide the reason(s) for your choice in part (a).
1
2
3
4
5
Jan
1030
896
951
991
1042
Feb
1032
793
861
950
997
Mar
1126
885
938
1024
1076
Apr
1285
1055
1109
1182
1198
May
1468
1204
1274
1349
1330
Jun
1637
1326
1422
1473
1407
Jul
1611
1303
1486
1540
1465
Aug
1608
1436
1555
1599
1516
Sep
1528
1473
1604
1591
1508
Oct
1420
1453
1600
1546
1431
Nov
1119
1170
1403
1316
1271
Dec
1013
1023
1209
1151
1083
Explanation / Answer
a) ARIMA ( auto regressive integrated moving average)
SVM(support vector machine)
b) the performance were evaluated based on three metrics : mean absolute error(MAE) ; mean absolute percentage error(MAPE); mean square error (MSE). The accuracy of statistical model in forecasting proved their effectiveness.
Although the comparisons found that no single method is completely superior to the others, the present study indeed highlighted that the SVM outperform the ARIMA model and decomposition methods in most cases.
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.