FORECASTI

NG
 A statement about the
future value of a
variable of interest.

The primary goal of

operations management
is to match supply to
demand.

Having a forecast of
FORECAST demand is essential for
determining how much
capacity or supply will be
needed to meet demand.
 Product Service
Design Design Finance

Operations
FORECAST
Accounting

MIS
Human
Marketing Resource
FORECAST ACCURACY
Forecast 𝑒 𝑡=
Error 𝐴𝑐𝑡𝑢𝑎𝑙𝑡 − 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 𝑡

Mean Absolute
Deviation
Mean Squared
Error
Mean Absolute
Percent Error
Types of
Forecasting
 QUALITATIVE FORECAST

Executive Sales Force Consumer Delphi

Opinion Opinions Surveys Technique
 STATISTICAL FORECAST
TIME SERIES
Time-ordered sequence of observations taken at regular intervals
(e.g.hourly, daily, weekly, monthly, quarterly, annually).
The data may be measurements of demand, earnings, profits,
shipments, accidents, output, precipitation, productivity, or the
consumer price index
 TIME SERIES – PATTERNS/BEHAVIOR

TREND SEASONAL
 TIME SERIES – PATTERNS/BEHAVIOR

CYCLE IRREGULAR
VARIATION
 TIME SERIES - NAÏVE METHOD
Naïve Forecast
uses a single previous value of a time series as the basis of a
forecast.

The naive approach can be

used with a stable series
(variations around an average),
with seasonal variations, or
with trend.
 TIME SERIES – MOVING AVERAGE
Moving Average
•Forecast is based on an average of recent values

∑ 𝐴𝑡 − 𝑖
𝐴 𝑡 − 𝑛 +…+ 𝐴𝑡 − 2 + 𝐴 𝑡 − 1
𝑖 =1
𝐹 𝑡 = 𝑀𝐴𝑛 = =
𝑛 𝑛

A weighted average is similar to a moving average, except

that it assigns more weight to the most recent values in a time
series.
 TIME SERIES – EXPONENTIAL SMOOTHING
•Exponential smoothing
a sophisticated weighted averaging method that is still relatively
easy to use and understand. Each new forecast is based on the
previous forecast plus a percentage of the difference between that
forecast and the actual value of the series at that point.

𝑁𝑒𝑥𝑡 𝐹𝑜𝑟𝑐𝑎𝑠𝑡=𝑃𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡+ 𝛼( 𝑃𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝐴𝑐𝑡𝑢𝑎𝑙 − 𝑃𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 )

Topics to Discuss
• Computation & Analysis • Sample Computation & Exercise
• Forecasting Accuracy
• Mean Absolute Deviation (MAD)
• Mean Squared Error (MSE)
• Mean Absolute Percent Error (MAPE)
• Forecasting Techniques
• Naïve Method
• Moving Average Method
• Weighted Moving Average Method
• Exponential Smoothing
• Time Series Regression
Forecasting Accuracy
Forecast 𝑒 𝑡=
Error 𝐴𝑐𝑡𝑢𝑎𝑙𝑡 − 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 𝑡

Mean Absolute
Deviation
Mean Squared
Error
Mean Absolute
Percent Error
Sample Computation
Compute for the
Forecast 𝑒 𝑡=
Error 𝐴𝑐𝑡𝑢𝑎𝑙𝑡 − 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡error
𝑡

𝑒1 Period 1
Period Actual Forecast e
𝑒1= 𝐴𝑐𝑡𝑢𝑎𝑙1 − 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 1
1 50 60 -10
𝑒 1=50 − 60
2 70 50 20
𝑒 1=−10
3 60 70 -10
4 80 60 20 𝑒2= 𝐴𝑐𝑡𝑢𝑎𝑙 2 − 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 2
5 50 80 -30 𝑒 2=70 − 50
𝑒 2=20
Sample Computation
Mean Absolute
Compute the
Deviation
absolute value of
error for each
Period Actual Forecast𝑒 |𝑒| period
|-10|
1 50 60 -10 10
¿ 𝑒 1∨¿ 10
2 70 50 20 20
3 60 70 -10 10 |20|
4 80 60 20 20 ¿ 𝑒 1∨¿ 20
Count
5 50 80 -30 30 n=5

Sum all |e| ∑ |𝑒𝑡|=¿

90 Compute MAD

𝑀 𝐴𝐷 =18
Sample Computation
Mean Squared
Error Compute the
squared value of
error for each
Period 𝑒|𝑒|
Actual Forecast 𝑒
2
period
1 50 60 -10 10 100
𝑒 21=100
2 70 50 20 20 400
3 60 70 -10 10 100
𝑒 22= 400
4 80 60 20 20 400
5 50 80 -30 30 900

Sum all ∑ 𝑒
2
=¿
1900 Compute MSE
 Compute the
Sample Computation absolute percent
error for each
Mean Absolute period
Percent Error
| 𝑒1
𝐴𝑐𝑡𝑢𝑎𝑙1 |
𝑥 100

| | 10
𝑥 100
Period 𝑒|𝑒|
Actual Forecast 𝑒
2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡 |
𝑥 100
50
0.20 𝑥 100
1 50 60 -10 10 100 20% %
2 70 50 20 20 400 29% | 27 00 | 𝑥 100
3 60 70 -10 10 100 17% 0.29 𝑥 100
4 80 60 20 20 400 25% %
5 50 80 -30 30 900 60%

Sum all ∑ | 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙 𝑡 |
𝑥 100=¿151%
Forecasting Techniques
Using this data, forecast the 6th
period with the least error
Period Actual

1 50
2 70
3 60
4 80
5 50

6
Naive 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 𝑡h𝑒6 𝑡h 𝑝𝑒𝑟𝑖𝑜𝑑
Impossible to Compute the
forecast. accuracy

𝐹 𝑡 = 𝐴𝑡 −1 No Actual on
previous period
Leave blank

Period Actual Forecast𝑒|𝑒| 𝑒

2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡 |
𝑥 100

1 50 - - - - -
2 70 50 20 20 400 29%
3 60 70 -10 10 100 17%
4 80 60 20 20 400 25%
5 50 80 -30 30 900 60%

6 50 80 1800 131%
 Moving Average
𝑛

∑ 𝐴𝑡 − 𝑖
𝐹 𝑡 = 𝑀𝐴𝑛 = 𝑖 =1 =
𝐴 𝑡 − 𝑛 +…+ 𝐴𝑡 − 2 + 𝐴 𝑡 − 1
To Forecast next period, average the 2
𝑛 𝑛
previous actual

Period Actual Forecast𝑒|𝑒| 𝑒

2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡 |
𝑥 100

1 50 Can’t Forecast. Need atleast two periods of previous actual

2 70 Can’t Forecast. Need atleast two periods of previous actual. Only 50 is given

3 60 60 0 0 0 0%
4 80 65 15 15 225 19%
5 50 70 -20 20 400 40%

6 65 35 625 59%
 Moving Average
𝑛

∑ 𝐴𝑡 − 𝑖
𝐹 𝑡 = 𝑀𝐴𝑛 = 𝑖 =1 =
𝐴 𝑡 − 𝑛 +…+ 𝐴𝑡 − 2 + 𝐴 𝑡 − 1
To Forecast next period, average the 3
𝑛 𝑛
previous actual

Period Actual Forecast𝑒|𝑒| 𝑒

2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡
𝑥 100 |
1 50 Can’t Forecast. Need atleast three periods of previous actual

2 70 Can’t Forecast. Need atleast three periods of previous actual. Only 50 is given

3 60 Can’t Forecast. Need atleast three periods of previous actual. Only 50 & 70 are given

4 80 60 20 20 400 25%
5 50 70 -20 20 400 40%

6 63.33 40 800 65%

 Weight Score
Weighted Moving Most
Average
Recent 2 Most Recent nd
Most Recent: 3
2nd Most Recent: 2
Note: Weight Score
The divisor should can be change
be the total weight
Period Actual score Forecast𝑒|𝑒| 𝑒
2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡 |
𝑥 100

2nd Most
Recent 1 50 Can’t Forecast. Need atleast two periods of previous actual
Most
Recent 2 70 Can’t Forecast. Need atleast two periods of previous actual. Only 50 is given
In POV of
Period 3 3 60 62 -2 2 4 3%
4 80 64 16 16 256 20%
5 50 72 -22 22 484 44%

6 62 40 744 67%
Exponential Smoothing
Alpha () can be set between 0-1
𝐹 𝑡 =𝐹 𝑡 −1 +𝛼 (𝑒 𝑡 −1 ) For this problem, alpha = 0.5
Since no Compute
previous error. To be
Period Actual
forecast, use
naïve method
use on
Forecast
formula
𝑒|𝑒| 𝑒
2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡 |
𝑥 100

1 50 Can’t Forecast. No previous Forecast. Leave Blank

2 70 50 20 20 400 29%
3 60 60 0 0 0 0%
4 80 60 20 20 400 25%
5 50 70 -20 20 400 40%

6 60 60 1200 94%
 Time Series Regression
𝐹 𝑡 =𝑎+ 𝑏𝑡 𝑏= 𝑛 ∑ 𝑡𝑦 − ∑ 𝑡 ∑ 𝑦 𝑎=
∑ 𝑦 −𝑏 ∑ 𝑡
𝑛∑ 𝑡 −¿¿¿ 𝑛
2

Period
(t)
Actual
(y)
𝑡𝑦 𝑡
2
Forecast𝑒|𝑒| 𝑒
2
| 𝑒𝑡
𝐴𝑐𝑡𝑢𝑎𝑙𝑡 |
𝑥 100

1 50 50 1 𝐹 1=59+1(1) 60 −1 0 10 1 00 %
2 70 140 4 𝐹 2=59+1 (2) 61 9 9 81 %
3 60 180 9 𝐹 3=59+ 1(3) 62 −2 2 4 3%
4 80 320 16 𝐹 4 =59+1( 4) 63 17 17 2 89 21%
5 50 250 25 𝐹 5=59+1 (5) 64 −1 4 14 1 96 28%

15 310 940 55 52 670 85%

5 ( 940 ) −(15)(310) 310 − 1 (15)
𝑏= 2 ¿1 𝑎=
5
¿ 59
5 (55)−(15)
𝐹 𝑡 =59+1 𝑡 @ 6 𝑡h 𝑝𝑒𝑟𝑖𝑜𝑑 𝐹 6=59 +1(6)¿ 65
Summary
Method MAD MSE MAPE Forecast
Naïve 20 600 32.75% 50
2 MA 11.67 312.5 19.67% 65
3MA 20 800 32.50% 63.33
2WMA(3,2) 13.33 372 22.33% 62
Exp (a=0.5) 15 400 23.50% 60
Time Series 10.4 167.5 17.00% 65

Check lowest for each accuracy

Summary
Method MAD MSE MAPE Forecast
Naïve 20 600 32.75% 50
2 MA 11.67 312.5 19.67% 65
3MA 20 800 32.50% 63.33
2WMA(3,2) 13.33 372 22.33% 62
Exp (a=0.5) 15 400 23.50% 60
Time Series 10.4 167.5 17.00% 65

6th Period Forecast is 65

Summary
Method MAD MSE MAPE Forecast
Naïve 20 600 32.75% 50
2 MA 11.67 312.5 19.67% 65
3MA 20 800 1.17% 63.33
2WMA(3,2) 13.33 372 22.33% 62
Exp (a=0.5) 15 150 23.50% 60
Time Series 10.4 167.5 17.00% 65

6th Period Forecast is 65

Questions?
Suggestions?
Feedback?
THANKYOU!!