Beruflich Dokumente
Kultur Dokumente
Nadia Khan
Application Areas
MEDICAL MILITARY TELECOM SCM MANAGEMENT FINANCE WEATHER POLITICS ASTRONOMY DEMOGRAPHY
8 Steps to Forecasting
1.Use or Objective? 2.What items/Quantities to be forecasted? 3.Time Horizon? 1 month (short term) 1 year (mid term) > 1 year (long term) 4.Select the forecasting Model 5.Gather Data 6.Validate the Forecasting Model 7.Make the forecast 8.Implement the results
Forecasting Techniques
Qualitative Models
Delphi Method Jury of Executive Opinion Sales Force Composite Consumer Market Survey
Causal Models
Regression Analysis Multiple Regression
Articial Intelligence
Articial Neural Networks Support Vendor Machines
Others
Simulation Prediction Market Probabilistic or Ensemble Forecasting Reference Class Forecasting
Qualitative Models
Decision Making Group
Delphi Method
Surveys/Questionnaire
Statistical Models
Qualitative Models
Nationwide Level Forecast Overall Forecast
Customer
Future Purchases
Forecasts
Forecast Errors
Measures of forecast accuracy include: Mean Absolute Deviation (MAD)
= |forecast errors| n
error actual n
100%
Forecast Accuracy
Example of Mobilink
Forecast Error Mean Absolute Deviation (MAD) Actual Value Forecast Value
Forecast Error n
Year
2005 2006 2007 2008 2009
|Actual-Forecast|
982900 1288350 5024565 14571238.5 18375950.775 40243004.275
On average each forecast missed the actual value by 8.04 Million Erls
8048600.8
Forecast Accuracy...Contd
Example of Mobilink
Forecast Error Mean Squared Error (MSE) Actual Value (Error) n
2
Forecast Value
Year
2005 2006 2007 2008 2009
(Error)
MSE n
1.15E+14
Forecast Accuracy...Contd
Example of Mobilink
Mean Absolute Percent Error (MAPE) (Error/actual) X 100 n
Year
2005 2006 2007 2008 2009
|(Error/actual)|
0.324614419234453 0.230769230769231 0.310344827586207 0.310344827586207 0.340329835082459 1.51640314025856
0.30X100
Forecast Accuracy...Contd
Example of Mobilink
Bias (Error) n
Bias --> not a good measure as negative errors can cancel out the positive errors
Year
2005 2006 2007 2008 2009
Error
982900 1288350 5024565 14571238.5 18375950.775 40243004.275
Bias = too high i.e. +8048600
8048600
Activity
Forecast Errors
Ms. Smith forecasted total hospital inpatient days last year. Now that the actual data are known, she is reevaluating her forecasting model. Compute the MAD, MSE, and MAPE for her forecast.
Month JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Forecast 250 320 275 260 250 275 300 325 320 350 365 380 Actual 243 315 286 256 241 298 292 333 326 378 382 396
Forecast Errors
Forecast JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC AVERAGE 250 320 275 260 250 275 300 325 320 350 365 380 Actual 243 315 286 256 241 298 292 333 326 378 382 396 |error| 7 5 11 4 9 23 8 8 6 28 17 16 MAD = 11.83 error^2 49 25 121 16 81 529 64 64 36 784 289 256 MSE = 192.83 |error/actual| 0.03 0.02 0.04 0.02 0.04 0.08 0.03 0.02 0.02 0.07 0.04 0.04 MAPE = .0368*100
3.68
Moving Averages
If all variations in a time series are due to random variations, with no trend, seasonal, or cyclical component, some type of averaging or smoothing model would be appropriate.
Moving Averages
Moving average methods consist of computing an average of the most recent n data values for the time series and using this average for the forecast of the next period.
Month Actual Shed Sales
10 12 13 16 19 23 26 (10+12+13)/3 = 11 2/3 (12+13+16)/3 = 13 2/3 (13+16+19)/3 = 16 (16+19+23)/3 = 19 1/3
January February
Weighted moving averages use weights to put more emphasis on certain recent periods. Month Actual Three-Month Weighted
d n) in
Shed Sales
January February March April May June July 10 12 13 16 19 23 26
pe rio dn w )( ei de gh m ts an d
pe
Moving Average
Weights Applied 3 2 1
fo r
gh t
(w
3*Sales last month + 2*Sales two months ago + 1*Sales three months ago
6 Sum of weights
ei
Exponential Smoothing
Exponential smoothing is a type of moving average technique that involves little record keeping of past data.
New forecast = previous forecast + (previous actual previous forecast) Mathematically this is expressed as:
Qtr Actual Tonnage Unloaded 180 168 159 175 190 205 180 182 Rounded Forecast using =0.10
175 176= 175.00+0.10(180-175) 175 =175.50+0.10(168-175.50) 173 =174.75+0.10(159-174.75) 173 =173.18+0.10(175-173.18) 175 =173.36+0.10(190-173.36) 178 =175.02+0.10(205-175.02) 178 =178.02+0.10(180-178.02) 179= 178.22+0.10(182-178.22)
the larger the smoothing parameter , the greater the weight given to the most recent value
2 3 4 5 6 7 8
Exponential Smoothing
Qtr Actual Tonnage Unloaded 180 168 159 175 190 205 180 182 Rounded Forecast using =0.50
Qtr
Actual Tonnage Unloaded 180 168 159 175 190 205 180 182
1 2 3 4 5 6 7 8
175 178 =175.00+0.50(180-175) 173 =177.50+0.50(168-177.50) 166 =172.75+0.50(159-172.75) 170 =165.88+0.50(175-165.88) 180 =170.44+0.50(190-170.44) 193 =180.22+0.50(205-180.22) 186 =192.61+0.50(180-192.61) 184 =186.30+0.50(182-186.30)
1 2 3 4 5 6 7 8
175 176= 175.00+0.10(180-175) 175 =175.50+0.10(168-175.50) 173 =174.75+0.10(159-174.75) 173 =173.18+0.10(175-173.18) 175 =173.36+0.10(190-173.36) 178 =175.02+0.10(205-175.02) 178 =178.02+0.10(180-178.02) 179= 178.22+0.10(182-178.22)
Exponential Smoothing
To select the best smoothing constant, evaluate the accuracy of each forecasting model.
Actual 180 168 159 175 190 205 180 182 Forecast with a = 0.10 175 176 175 173 173 175 178 178 Absolute Deviations 5 8 16 2 17 30 2 4 Forecast with a = 0.50 175 178 173 166 170 180 193 186 Absolute Deviations 5 10 14 9 20 25 13 4
Activity
Class Example
PM Computer assembles customized personal computers from generic parts. The owners purchase generic computer parts in volume at a discount from a variety of sources whenever they see a good deal. It is important that they develop a good forecast of demand for their computers so they can purchase component parts efficiently.
Period 1 2 3 4 5 6 7 8 9 month Jan Feb Mar Apr May June July Aug Sept actual demand 37 40 41 37 45 50 43 47 56
Compute a 2-month moving average Compute a 3-month weighted average using weights of 4,2,1 for the past three months of data Compute an exponential smoothing forecast using = 0.7 Using MAD, what forecast is most accurate?
MAD