Forecasting

Forecasting can be described as the act of predicting future events.

Techniques used in forecasting:

1. Naive approach
2. Moving average
3. Exponential smoothing 
4. Trend projection
5. Linear regression
6. Coefficient correlation

Importance of forecasting:

• Better understand seasonal peaks and troughs.
• Can forecast sales.
• Will know when to order inventory.
• What is the best time to launch a new product.
• Right time to develop a product.
• Whether firm needs additional staff.(when can afford)
• Estimate financial requirements.
• Make better management decisions.

Naive Approach

Assumes that the demand in the next period is equal to the demand of the previous period.

Moving Average

In statistics, a moving average is a calculation to analyze data points by creating series of averages of different subsets of the full data set.

Example: The sales of cars for a certain company are shown in the table below. Development forecast for week 4 to 7 using a 3 week moving average approach.

Exponential smoothing

Exponential smoothing is a rule of thumb technique for smoothing time series data, particularly for recursively applying as many as three low-pass filters with exponential window functions. Such techniques have broad application that is not intended to be strictly accurate or reliable for every situation.

Ft = Ft-1 + α(At-1 -Ft-1)

Ft : New forecast

Ft-1: Forecast of the previous period.

At-1: Actual value or actual demand of the previous period.

α: Smoothing constant (0 ≤ α ≤ 1)

Example: Use exponential smoothing model with a smoothing constant of 0.6 to forecast demand from the information given in the table below.

Mean Absolute deviation (MAD)

The Mean absolute deviation of a data set is the average of the absolute deviations from a central point.

It is a measure of overall forecast used to evaluate the accuracy of different forecasting models.

The lower the MAD, the more accurate is the overall forecasting model.

Absolute deviation = |A - F|

From previous example: |500 - 490| = 10, |550 - 496| = 9, 8.6, 11.6, 9.6, 16.2

MAD = (10 + 9 + 8.6 + 11.6 + 9.6 + 16.2)/6 = 10.8

Trend projection

When numerical data are available, a trend can be plotted on graph paper to show changes through time. Formula: y = bx + a

Coefficient of correlation (r)

Measure of strength between two variables.

University of Mauritius past papers Question and Answer

May 2014 Question 4
(a)  Explain the importance of forecasting to the different areas of operations in an organization.  [5 marks]
       In notes (TOP ↑)

 (b)  Sales of refrigerators in a company have grown steadily during the past five years as shown in the table hereunder :
The sales manager had predicted, before the business started, that the sales in Year 1 would be 410 refrigerators.
(i)  Using smoothing constants of 0.6 and 0.9, develop a forecast for the sales of refrigerators from Year 2 up to Year 6, both years inclusive.[5 marks]

(ii) Determine which smoothing constant gives the most accurate forecast.[4 marks]
MAD(0.6): [|450-410| + |495-434| + 47.04 + 63.96 + 46.58]/5 =51.716
MAD(0.9): [40 + 49 + 27.99 + 56.16 + 26.62]/5 = 39.95
Smoothing constant (0.9) is the most accurate forecast.

(c) Bus and subway ridership for the summer months in London is believed to be tied heavily to the number of tourists visiting the city.  During the past 12 years, the following data have been obtained :
(i) Develop a regression relationship.[5 marks]
(ii) What is the expected ridership if 10 million tourists visit London in a year? [2 marks]      
(iii) Compute the coefficient of correlation.  [4 marks]