The Importance of Statistical Models in Forecasting

Unit VIII Final Project
Author’s Name
Institutional Affiliation
Table of Content
TOC \o "1-3" \h \z \u 1. Introduction PAGEREF _Toc31903201 \h 32. Forecast Accuracy PAGEREF _Toc31903202 \h 33. Impact of Differences Between the Actual and Forecasted Values for the Town PAGEREF _Toc31903203 \h 44. An explanation for Selection of Methodology PAGEREF _Toc31903204 \h 55. New Forecasting Model for High Accuracy PAGEREF _Toc31903205 \h 56. Conclusion PAGEREF _Toc31903206 \h 6References PAGEREF _Toc31903207 \h 8
UNIT VIII FINAL PROJECT
1. Introduction
Optimized results can be achieved with the use of statistical models for the purpose of forecasting. Ineffective forecasting models result in results with errors. Therefore, the proper analysis needs to be conducted before the selection of appropriate statistical models and then forecasted values need to be compared with the actual values to identify the differences and make changes in the current models to ensure that errors are minimized. The report presents an analysis of the forecasted and actual results of hurricanes by a regional weather office on the Atlantic coast. Monica, an employee at the regional office has identified that there exists a difference between the actual and forecasted results of the observed landfall. Therefore, it becomes essential to analyze the difference, its impact on the town and then recommend a more effective method for the purpose of forecasting.
2. Forecast Accuracy
In order to analyze the accuracy of the forecasted results, the selected methods are mean absolute deviation (MAD), mean squared error (MSE) and mean absolute percent error (MAPE) (Render, Stair, Hanna & Hale, 2015). The method of MAD is used to analyze and measure the spread in the data set, and it is considered as one of the best tools in the situation where the data is normal. MSE is a method used by calculations of the averages of error squares. MAPE is another statistical method used for the evaluation of the effectiveness of the forecasting system and the method is effective in the situation, where there are no extreme values in the data sets or the number of zeros is less. Although for the current case scenario, MAPE is selected for the purpose of analysis to evaluate whether the current forecasting system used by the regional office is an effective one to predict the landfall of the hurricanes.
Figure 1: Calculated Mean Absolute Deviation (MAD), Mean Squared Error (MSE) and Mean Absolute Percent Error (MAPE)
Years

Actual

24th-hour Forecast

The absolute value of errors (Deviation), (Actual – Forecast)

Mean Squared Error
Errors Error square

Mean absolute percent error

1

4

6

| 4-6 | = 2

-2

4

50%

2

5

5

| 5-5 | = 0

0

0

0%

3

30

40

|30-40| = 10

-10

100

33.33%

4

2

10

| 2-10 | = 8

-8

64

400%

5

12

13

|12-13| = 1

-1

1

8.33%

6

7

5

| 7-5 | = 2

2

4