Forecasting
Model
Forecasting: Some examples & context
Manufacturing
A manufacturer of household appliances wants to add to
add another product line for manufacturing microwave
ovens The decision requires a good understanding of the
nature of demand for the range of microwave ovens
proposed to be manufactured
Services
A hospital chooses to add one more specialty health care
wing, it needs to make some assumptions about the demand
for the facility
Public Policy
Government of India needs to have a reasonable estimate of
the population growth over the next 10 – 20 years while it
formulates long term plans for creating infrastructure for
transport
Forecasting
Predicting the Future
Forecasts are estimates of
magnitude and
timing
of uncertain events that happen
in every business setting
Two type of Forecasting
Method
Qualitative forecast methods
subjective
Quantitative forecast methods
based on mathematical formulas
Forecasting: Time Horizon
Criterion Short-term Medium-term Long-term
Typical Duration 1-3 months 12-18 months 5-10 years
Nature of Purely Tactical Tactical as well as Purely Strategic
decisions Strategic
Key Random (Short Seasonal and Long-term trend
considerations Term) effects cyclical effects business cycle
Nature of data Mostly Subjective & Largely subjective
quantitative Quantitative
Degree of Low Significant High
Uncertainty
Some examples Revising Annual New Product
quarterly production Introduction
production plans Planning Facilities location
Rescheduling Capacity decisions
supply of raw Augmentation New business
material development
What to forecast
Car Manufacturing Company :
Independent and Dependent Demand
Independent Demand : Dependent Demand : Raw
final Product of Company Materials and components of
final product of Company
To Predict the future of
Independent Demand : Use
Forecasting
For Dependent Demand : Use
MRP/ ERP
FORECASTING (FORECASTING PROCESS: (IS CONTINUOUS)
IDENTIFY THE PURPOSE
COLLECT DATA
PLOT AND IDENTIFY PATTERNS
SELECT A MODEL (APPROPRIATE)
COMPUTE FORECAST FOR THE
PERIOD OF HISTORICAL DATA SELECT
NEW
MODEL
CHECK FORECAST ACCURACY
or
MODIFY
IS ACCURACY No
ACCEPTABLE?
YES
FORECAST OVER THE
PLANNING PERIOD
ADJUST FORECAST WITH ADDITIONAL
QUALITATIVE INFORMATION
MONITOR RESULTS AND MEASURE ACCURACY
QUALITATIVE Models for
forecasting
Subjective Judgment Methods
Draw substantially from the expertise of a
group of senior managers using some
collective decision making framework
Examples
Delphi Method
Executive Judgment
Market Research
QUANTITATIVE Models for
Forecasting
Extrapolative Causal or
Models Explanatory Models
Make use of past data and Analyse the data from a
essentially prepare the point of cause – effect
future estimate by some relationship
method of extrapolating the Examples
past data Multiple Regression
Models
Examples
Econometric Models
Moving Averages –
Weighted, Simple Technological
Forecasting
Exponential Smoothening
Time Series Methods
Forecasting Methods
M1: Simple Moving Average forecasting
Model
M2: Weighted Moving Average Forecasting
Model
M3: Exponential Smoothing Forecasting
Model
M4: Time Series Analysis (Linear Trend Line)
M5: Time Series analysis with Seasonality
Simple Moving Average forecasting Model
The simple moving average model assumes
an average is a good estimator of future
behavior
The formula for the simple moving average
is:
A t-1 + A t-2 + A t-3 +...+A t- n
Ft =
n
Ft = Forecast for the coming period
N = Number of periods to be averaged
A t-1 = Actual occurrence in the past period for up to “n”
periods
Simple Moving Average Example
A t-1 + A t-2 + A t-3 +...+A t- n
Month Demand Ft = n
Jan 650
Feb 678
March 720 • What are the 3-month
April 785 moving average forecasts
May 859 for demand of August?
June 920
July 850
Solution
Month Demand Forecasting
Moving Average forecasting
Jan. 650
April = (720+678+650)/3 = 682.67
Feb 678
March 720 May = (785+720 +678)/3 = 727.67
April 785 682.67 June = (859+785+720)/3 = 788
May 859
727.67 July = (920+859+785)/3 = 854.67
June 920 788
July 850
August = (850+920+859)/3 =
854.67
876.33
August 876.33
Simple Moving Average Problem (1)
A t-1 + A t-2 + A t-3 +...+A t- n
Ft =
Week Demand n
1 650
2 678
3 720 Question: What are the
4 785 3-week and 6-week
5
6
859
920
moving average
7 850 forecasts for demand?
8 758
9 892
10 920
11 789
12 844
Solution
3-week moving average forecasts for demand
F13 = (D12 +D11 +D10)/3 = (844+879+920)/3 = 851
6-week moving average forecasts for demand
F13 = (D12 +D11 +D10 + D9 +D8 +D7)/6
(844+879+920+892+758+850)/6 = 842.17
17
Calculating the moving averages gives us:
Week Demand 3-Week 6-Week
1 650 F4=(650+678+720)/3
2 678 =682.67
3 720 F7=(650+678+720
4 785 682.67 +785+859+920)/6
5 859 727.67 =768.67
6 920 788.00
7 850 854.67 768.67
8 758 876.33 802.00
9 892 842.67 815.33
10 920 833.33 844.00
11 789 856.67 866.50
12 844 867.00 854.83
Simple Moving Average Problem (2)
Week Demand
1 820
What is the 3 week
2 775 moving average
3 680 forecast for 8th
4 655 Week?
5 620
6 600
7 575
Solution
3-week moving average forecasts for
demand
F8 = (D7 +D6 +D5)/3 = (575 + 600 + 620)/3
= 598.33
Simple Moving Average Problem (2) Solution
Week Demand 3-Week 5-Week
1 820 F4=(820+775+680)/3
2 775 =758.33
3 680 F6=(820+775+680
+655+620)/5
4 655 758.33 =710.00
5 620 703.33
6 600 651.67 710.00
7 575 625.00 666.00
Weighted Moving Average
Forecasting Model
While the moving average formula implies an equal
weight being placed on each value that is being averaged,
the weighted moving average permits an unequal
weighting on prior time periods
The formula for the moving average is:
Ft = w 1 A t -1 + w 2 A t - 2 + w 3 A t -3 + ...+ w n A t - n
n
wt = weight given to time period “t”
occurrence (weights must add to one)
w
i=1
i =1
Weighted Moving Average Problem (1)
Question: Given the weekly demand and weights, what is
the forecast for the 4th period or Week 4?
Week Demand Weights:
1 650
2 678 Week 3 0.5
3 720 Week 2 0.3
4 Week 1 0.2
Note that the weights place more emphasis on the
most recent data, that is time period “t-1”
Weighted Moving Average Problem (1)
Solution
Week Demand Forecast
1 650
2 678
3 720
4 693.4
F4 = 0.5(720)+0.3(678)+0.2(650)=693.4
Weighted Moving Average Problem (2)
Question: Given the weekly demand information and
weights, what is the weighted moving average forecast
of the 5th period or week?
Week Demand Weights:
1 820 Week 4 0.7
2 775 Week 3 0.2
3 680
Week 2 0.1
4 655
Weighted Moving Average Problem (2)
Solution
Week Demand Forecast
1 820
2 775
3 680
4 655
5 672
F5 = (0.1)(755)+(0.2)(680)+(0.7)(655)= 672
Disadvantage for moving
average method
It does not react to variations that occur for
a reason, such as cycle and seasonal
effects.
It is the need to continually carry a large
amount of historical data.
Exponential Smoothing Forecasting Model
EXPONENTIAL SMOOTHING IS ALSO AN AVERAGING
METHOD THAT WEIGHTS THE MOST RECENT DATA MORE
STRONGLY. THE FORECAST WILL REACT MORE TO RECENT
CHANGES IN DEMAND.
EXPONENTIAL SMOOTHING (POPULAR / FREQUENTLY
USED) - REQUIRES MINIMAL DATA - EASY TO
UNDERSTAND - COMPUTER SOFTWARE
Ft +1 = Dt + (1 − )Ft ;
Ft+1 = Forecast for the next period, Dt = actual demand at
present
Ft = previously determined forecast for present period
= weighting factor (smoothing constant) (o 1)
Exponential Smoothing
The smoothing constant, , must be
between 0.0 and 1.0.
A large provides a high impulse response
forecast.
A small provides a low impulse response
forecast.
FORECASTING
EXPONENTIAL SMOOTHING FORECASTS, = 0.3
FORECAST FT+1
PERIOD MONTH DEMAND = 0.3
1. JANUARY 40 -
2. FEBRUARY 41
3. MARCH 37
4. APRIL 45
5. MAY 50
6. JUNE 43
7. JULY 47
8. AUGUST 56
9. SEPTEMBER 52
10. OCTOBER 55
11. NOVEMBER -
F2 = . D1 + (1- ) F1 = (.3) (40) + (.7) (40) = 40
F3 = D2 + (1- ) F2 = (.3) (41) + (.7) (40) = 40.3
EXPONENTIAL SMOOTHING FORECASTS, = 0.3
Period Month Demand (D) Forecasting (α = 0.3)
1 January 40 F1 = D1 = 40
2 February 41 F2 = (40*0.3)+ (40*0.7) = 40
F3 = (41*0.3)+ (40*0.7) = 40.30
3 March 37
4 April 45 F4 = (37*0.3)+ (40.30*0.7) = 39.30
5 May 50 F5 = (45*0.3)+ (39.30*0.7) = 41.01
6 June 43 F6 = (50*0.3)+ (41.01*0.7) = 43.70
7 July 47 F7 = (43*0.3)+ (43.70*0.7) = 43.49
8 August 56 F8 = (47*0.3)+ (43.49*0.7) = 44.54
9 September 52 F9 = (56*0.3)+ (44.54*0.7) = 47.97
10 October 55 F10 = (52*0.3)+ (47.97*0.7) = 49.18
11 November F11 = (55*0.3)+ (49.18*0.7) = 50.92
Example
Indian Bakers are specialists in cakes,
pastries, and other bakery products in
the city of Jaipur. The actual demand Week Actual
(in units) experienced by them for their No. Demand
popular birthday cakes during the past (units)
7 weeks is given in Table.
1 50
Find the forecasting for 8th week by:
a. Forecast the demand by three and six week 2 62
Moving Average Method 3 66
Forecast the demand by Weight Moving
b.
4 74
Average Method if weights of 0.333 on
the present period, 0.25 one period ago, 5 80
0.25 two periods ago and 0.167 three
6 84
periods ago.
c. Forecast the demand by Exponential 7 78
Smoothing Method if the smoothing
constant (α) is 0.70.
Components of Demand
Trend
Seasonal element
Cyclical elements
Random variation
Demand Behavior
Trend
a gradual, long-term up or down movement of
demand
Random variations
movements in demand that do not follow a pattern
Cycle
an up-and-down repetitive movement in demand
Seasonal pattern
an up-and-down repetitive movement in demand
occurring periodically
Forms of Forecast Movement
Demand
Demand
Random
movement
Time Time
(a) Trend (b) Cycle
Demand
Demand
Time Time
(c) Seasonal pattern (d) Trend with seasonal pattern
Linear Trend Line
xy - nxy
y = a + bx b =
x2 - nx2
where a = y-bx
a = intercept
b = slope of the line where
n = number of periods
x = time period
y = forecast for x
demand for period x x = = mean of the x values
n
y
y = = mean of the y values
n
Example
AirLite Inc. manufacturer’s high-quality tennis racquets. The
production manager wants to develop a forecasting system
to use for future production resource planning. Yearly
demand data for the past four years is given below:
Develop forecast levels for next three years (2024, 2025,
2026).
Year Demand (Lakhs of racquets)
2019 15
2020 20
2021 24
2022 19
2023 28
ẋ = 15/5 = 3
Year x Demand (y) xy x^2 ẏ = 106/5 = 21.2
b = (343 –
2019 1 15 15 1
(5*3*21.2))/ (55-
2020 2 20 40 4
5(3)^2) = 2.5
2021 3 24 72 9
a = 21.2 – (2.5*3) =
2022 4 19 76 16
13.7
2023 5 28 140 25
Sum 15 106 343 55 y = 13.7 + 2.5 x
Year x Forecasting (Lakhs of Racquets)
2024 6 = 13.7 + (2.5*6) =28.7
2025 7 =13.7 + (2.5*7) = 31.2
2026 8 = 13.7 +(2.5*8) = 33.7
Seasonality in time series
forecasting
It is usually fluctuations that take place
within one year and tend to be repeated
annually.
These seasons can be caused or determined
by weather, holidays, or other phenomena.
Seasonal pattern is a repetitive increase and
decrease in demand.
Exam.: retail items, including toys, sports
equipment, clothing, electronic appliances,
etc.
SEASONAL ADJUSTMENTS
A SEASONAL FACTOR IS A NUMERICAL VALUE
THAT IS MULTIPLIED BY THE NORMAL
FORECAST TO GET A SEASONALLY ADJUSTED
FORECAST.
O Si 1
D1
S1 =
D
Example
Wishbone farms grow turkeys to sell to a meat-processing company
throughout the year. However, its peak season is obviously during the
fourth quarter of the year, from October to December. Wishbone
farms has experienced the demand for turkeys for the past three years
shown in the following table:
DEMAND (1000s) FOR QUARTER
YEAR 1 2 3 4 TOTAL
2021 12.6 8.6 6.3 17.5 45.0
2022 14.1 10.3 7.5 18.2 50.1
2023 15.3 10.6 8.1 19.6 53.6
TOTAL 42.0 29.5 21.9 55.3 148.7
We can compute the seasonal factor by
dividing total quarterly demand for three
years by total demand across all the
three years.
S1 = 42/148.7 = 0.28
S2 = 29.5 / 148.7 = 0.20
S3 = 21.9 / 148.7 = 0.15
S4 = 55.3 / 148.7 = 0.37
Next, we want to multiply the forecasting
demand for the next year, 2024, by each of
the seasonal factors to get the forecasting
demand for each quarter. For this, we need a
demand forecast for 2024.
Table show increasing trend. We compute a
linear trend line for the three years of data .
Trend line: y = 40.97 + 4.30 x
Forecast for 2024 = 40.97 + 4.30(4) = 58.17
Thus the forecast for 2024 is 58.17 or 58170
turkeys.
The seasonally adjusted forecasts, SF,
for 2024 are
SF1 = (S1)(F) = 0.28*58.17 = 16.28
SF2 = 0.20*58.17 = 11.63
SF3 = 0.15*58.17 = 8.73
SF4 = 0.37*58.17 = 21.53
FORECASTING Error
FORECAST ACCURACY
FORECAST ERROR IS THE DIFFERENCE BETWEEN THE
FORECAST AND THE ACTUAL DEMAND
MEASURES OF FORECAST ERROR
MEAN ABSOLUTE DEVIATION (MAD)
MEAN ABSOLUTE PERCENT DEVIATION (MAPD)
(MAPE)
FORECASTING
MAD IS THE AVERAGE, ABSOLUTE DIFFERENCE
BETWEEN THE FORECAST AND DEMAND
MAD = | Dt - Ft | ;
n
WHERE T = PERIOD NUMBER
Dt = DEMAND IN PERIOD T
Ft = FORECAST FOR PERIOD T
n = TOTAL NO. OF PERIODS.
FORECASTING
FORECAST ACCURACY
PERIOD DEMAND (Dt) FORECAST ERROR | Dt - Ft|
Ft( = .3) Dt - Ft
1. 40 - -
2. 41 40 1.00 1.00
3. 37 40.30 -3.30 3.30
4. 45 39.31 5.69 5.69
5. 50 41.02 8.98 8.98
6. 43 43.71 -0.71 0.71
7. 47 43.50 3.50 3.50
8. 56 44.55 1.45 1.45
SUMMARY OF KEY FORMULAS
Moving Average Adjusted Exponential AFt+1 =Ft+1 +Tt+1
n Smoothing
∑ Di
MAn= i= 1 z
n
Weighted Moving Trend Factor Tt+1 =β(Ft+1 - Ft) + (1-β) Tt
Average n
∑ Wi Di
WMAn= i= 1 z Linear Trend Line γ= a + bx
Exponential Smoothing F t+1 = αDt + (1- α) Ft Least Squares b= ∑xγ – nxγ
∑x2 – nx2
a= y - bx
