MGT 304

Quantitative Methods
for Business
Dr. Thamer Almutairi
 Topic 6
Forecasting
¬ Explain the difference between the forecast approaches. (Qualitative vs.

Quantitative)

¬ Describe time series models.

¬ Apply the naive, moving-average, exponential smoothing, and trend methods.

¬ Develop seasonal indices.

¬ Compute three measures of forecast accuracy.

Forecastin
g
What is forecasting ? Process of predicting a future event.

• With the uncertain environment, Managers are always trying to make better
estimates of what will happen in the future.
® Production ® Personnel
• Underlying basis of all business decisions.
® Inventory ® Facilities

Forecast Approaches

i. Qualitative forecasts: use a variety of factors such as the decision maker’s

intuition, emotions, and personal experiences.

ii. Quantitative forecasts: use a variety of mathematical models that rely on

historical data and/or associative variables to forecast demand.
I. Qualitative Forecasts:
Disadvantage
:
1. Jury of executive opinion could be “groupthink”
pool opinions of a small group of high-level experts &
managers.
2. Delphi method take long time no
consensus
an iterative, written process that uses a panel of experts.

3. Sales force composite misrepresent the truth for

incentives
estimates from individual salespersons are reviewed, then
aggregated.
4. Consumer market survey the sample is
bias
ask the customer (market
research).
II. Quantitative forecasts: “using historical data”

1. Naive Approach

2. Moving Average Time-Series Models

3. Weighted Moving Average
look at what has happened over a period of
4. Exponential Smoothing time and use a series of past data to make a
forecast.
5. Trend Projection
Associative Model

6. Linear Regression incorporate the variables or factors that

might influence the quantity being forecast.
Types of Forecasts
Organizations use three major types of forecasts in planning future operations:

o Economic
forecast
is the process of attempting to predict the future condition of the economy (e.g.,
interest rates)
o Technological
forecast
is the process of predicting technological change which can result in the birth of
new product. (predict new product sales)

o Demand forecasts
are projections of demand for a company’s products or services. (predict existing
product sales)
Time-Series Forecasting: A time series is based on a sequence of evenly spaced
(weekly, monthly, quarterly) data points.

Decomposition of a Time Series:

1. Random variations: are no clear patterns in the
2. data.
Trend: is the gradual upward or downward movement of the data over
3. time.
Seasonality: is a data pattern that repeats itself after a period of time.
Period “Season” # of “Seasons” In
Length Length Pattern
Week Day 7
Month Week 4 – 4.5
Month Day 28 – 31
Year Quarter 4
Year Month 12
Year Week 52

4. Cycles: are patterns in the data that occur every several

years.
 Trend
componen
t
Demand for product or service Seasonal peaks

Actual
demand line

Average
demand over 4
years
Random variation
| | | |
1 2 3 4
Time (years)
1. Naive
The simplest way to forecast is to assume that demand in the next period will be
approach
equal to demand in the most recent period.

2. Moving Average Method

Uses several historical actual data values to generate a forecast.
“Useful if we can assume that market demands will stay fairly steady over
time”.

Moving Avrage =
∑ demand in previous n period
n
Example 1: You’re the manager of a museum store that sells historical replicas. You
want to forecast sales for June using a 3-period moving average.

Month Actual Sales

January 4,000

February 6,000

March 6,000

April 4,000

May 8,000 Forecast Sales

June
3. Weighted Moving Average

Method
® Used when trend is present.
® Older data usually less important.
® Weights based on intuition.
® Weights often between 0 and 1 and sum to
1.

Weighted Moving Avrage =

∑ ( ( weighted for period n ) × ( demand in period n ) )
∑ weights
Example 2: Sales of electric coffee makers at a local retail store over the last five
months are shown below. Using weights of 1, 2, 3, and 4, prepare a forecast for
June. More recent data has the higher weights.
Month Actual Sales Weight
January 90 0
February 70 1
March 80 2
April 85 3
May 82 4
Example 2: Sales of electric coffee makers at a local retail store over the last five
months are shown below. Using weights of 1, 2, 3, and 4, prepare a forecast for
June. More recent data has the higher weights.
Weight
Month Actual Sales Weight
(%)
January 90 0
February 70 1
March 80 2
April 85 3
May 82 4
Example 3:

Year 1 2 3 4 5 6 7 8 9 10 11
Demand 7 9 5 9 13 8 12 13 9 11 7

a. Plot the above data on a graph. Do you observe any trend, cycles, or random
variations?
b. Starting in year 4 and going to year 12, forecast demand using a 3-year moving
average. Plot your forecast on the same graph as the original data.
c. Starting in year 4 and going to year 12, forecast demand using a 3-year moving
average with weights of .1, .3, and .6, using .6 for the most recent year. Plot this
forecast on the same graph.
d. As you compare forecasts with the original data, which seems to give the better
results?
4. Exponential Smoothing Method

o Form of weighted moving average.

• Weights decline exponentially.

weight
• Most recent data weighted most.
never zero
o Requires smoothing constant (α).

s
• Ranges from 0 to 1. age of
recen old
data
• Subjectively chosen. t

o Involves little record keeping of past data.

Exponential Smoothing
The latest estimate of demand is equal to the
FEquation
t = Forecast value for period t
old forecast adjusted by a fraction of the
A t = Demand at period t
difference between the last period’s actual
α = Smoothing constant demand and last period's forecast

Forecast for new period is:

New Forecast = Last period’s forecast + α ( Last period’s actual demand – Last period’s
forecast)

Ft+1 = Ft + (α (At − Ft))

Example 4: You’re organizing a meeting. You want to forecast attendance for your
6 using exponential smoothing . The year 1 forecast was 1750.

Year Demand Forecast,

1 1800 1750

2 1680

3 1590

4 1750

5 2000

6 NA
Forecasting effects of the smoothing constant

The exponential smoothing formula can also be written as:

If is Prior period 2 periods ago 3 periods ago

10% 10% 9% 8.1%
50% 50% 25% 12.5%
90% 90% 9% 0.9%

Implication:
• Choose high values of α when underlying average is likely to change
• Choose low values of α when underlying average is stable
Cont. Example 3:

Year 1 2 3 4 5 6 7 8 9 10 11
Demand 7 9 5 9 13 8 12 13 9 11 7

Develop a forecast for years 2 through 12 using exponential smoothing with and a
forecast for year 1 of 6. Plot your new forecast on a graph with the actual data and
the naive forecast. Based on a visual inspection, which forecast is better?
Regression is useful for two types of forecasting: time series and causal

4. Trend Projection 5. Linear

y^ = a + b( t ) y^ = a + b( x )
Regression

y^ : dependent variale ( predicted value) x

a : y axis intercept
b: slope of the regression line
t : time

a = y− bx b = ∑ xy − n x y
∑ x
2
− n x
2
Example 5: Demand for the last six months was 25, 23, 30, 34, 38, and 40,
respectively. Using linear regression, make a forecast for the next three months.

Month Demand
1 25
2 23
3 30
4 34
5 38
6 40
Seasonality
Seasonal variations in data are regular movements in a time series that relate to
recurring events such as weather or holidays.

Forecasting with seasonal data:

1. Compute a seasonal index for each season by dividing that season’s historical
average demand by the average demand over all seasons.
2. Estimate next year’s total annual demand.
3. Divide this estimate of total annual demand by the number of seasons, then
multiply it by the seasonal index for each season.

This provides the seasonal forecast.

Example 6: Forecasting with Seasonality. "Forecast for next year =
10,400"
Average demand over the last 5 Average Seasonal Demand
years: = (2000 + 3200 + 2400 + 1600) / 4
Spring: 2000 Fall: 2400 = 2300
Winter: 1600 Summer: 3200
Example 7: What will be the forecast level for the product in the Spring of
year 5?
Se Yea
as o r 1 2 3 4 5
n
Winter 1400 1200 1000 900
Spring 1500 1400 1600 1900
Summer 1000 2100 2000 1500
Fall 600 750 650 500
Total 4500 5450 5250 4800 6000

• Using a different method • Average over all seasons

Example 7: What will be the forecast level for the product in the Spring
of year 5? Y
Se e
as o ar 1 2 3 4 5
n
Winter 1400 1200 1000 900
Spring 1500 1400 1600 1900 1920
Summer 1000 2100 2000 1500
Fall 600 750 650 500
Total 4500 5450 5250 4800 6000

• Average over Spring • Spring Index • Forecast for Spring of

year 5
Example 7: What will be the forecast level for the product in the Spring
of year 5?
Se Yea
as o r 1 2 3 4 5
n
Winter 1400 1200 1000 900 1350
Spring 1500 1400 1600 1900 1920
Summer 1000 2100 2000 1500
Fall 600 750 650 500
Total 4500 5450 5250 4800 6000

• Average over Winter • Winter Index • Forecast for Winter of year

5
Example 7: What will be the forecast level for the product in the Spring
of year 5?
Se Yea
as o r 1 2 3 4 5
n
Winter 1400 1200 1000 900 1350
Spring 1500 1400 1600 1900 1920
Summer 1000 2100 2000 1500 1980
Fall 600 750 650 500
Total 4500 5450 5250 4800 6000

• Average over Summer • Summer Index • Forecast for Summer of

year 5
Example 7: What will be the forecast level for the product in the Spring
of year 5?
Se Yea
as o r 1 2 3 4 5
n
Winter 1400 1200 1000 900 1350
Spring 1500 1400 1600 1900 1920
Summer 1000 2100 2000 1500 1980
Fall 600 750 650 500 750
Total 4500 5450 5250 4800 6000

• Average over Fall • Fall Index • Forecast for Fall of year

5
Measuring Forecast Error

Forecast error = Actual demand − Forecast value Forecast error = A t − F t

o Several measures are used to calculate the over all forecast

error
o These measures can be used to compare different forecast

i. models
Mean Squared Error (MSE)
n

∑ Forecast
error
2 ∑ ( At − F t )2
MSE = = t =1
Number of forecasts n
Measuring Forecast Error

Forecast error = Actual demand − Forecast value Forecast error = A t − F t

o Several measures are used to calculate the over all forecast

error
o These measures can be used to compare different forecast

ii. models
Mean Absolute Deviation (MAD)
n

∑ | Forecast error |
∑ | At − Ft |
t =1
MAD = =
Number of forecasts n
Measuring Forecast Error

Forecast error = Actual demand − Forecast value Forecast error = A t − F t

o Several measures are used to calculate the over all forecast

error
o These measures can be used to compare different forecast

iii.models
Mean Absolute Percent Error (MAPE)
Example 8: You’re a marketing analyst for Hasbro Toys. You’ve forecast sales (in
thousands) with a linear model and an exponential smoothing model with α = 0.9.
Which model should you use?
Actual Forecast

Year Sales Linear Model Exponential Smoothing

1 1 0.6 1

2 1 1.3 1

3 2 2 1

4 2 2.7 1.9

5 4 3.4 2
Example 9: The table below provides actual sales for years 1 through 7. The firm
uses a three-year moving average to make forecasts.

a. What is the forecast for year

8?
Forecas
Year Sales
t
1 100
2 200
3 300 b. What is the MAD based on these data?
4 400
5 100
6 900
7
8 800
Thank You