There are different Forecasting techniques that being used for a maturely stable historical data. These forecasting method is really helpful to give us a indication if the sales is going in a upward or downward trend. Let us observe the mathematical logic behind below models.
1.Simple Average
2.Simple Moving Average
3.Weighted Average
4.Weighted Moving Average
Simple Average: In this algorithm, forecast is equal to the Average of historical data of N period. N is equal to Historical Period.
Information maintained in Forecast Model profile
Output after successful completion of application job.
The forecast horizon is maintain as 12 Month and there are 12 historical data points. May 2021 is the current month so before that will have ex-post forecast.
You can download the file for the logic of these stat model.
https://drive.google.com/file/d/10pGnn8spX76MtQtVQ5dbi4Cv5XZzSARB/view?usp=sharing
Simple Moving Average:
Simple Moving Average algorithm is maintained in Mange Forecast Model App.
There are two option in Simple Moving Average algorithm where in one future forecast will remain constant and in another with the extended calculation horizon checked on, system takes historical data plus the previous generated forecast number.
Below is the information maintained without extended calculation horizon
Below is the information maintained with extended calculation horizon
Output after successful completion of application job.
Extended Calculation ON
The forecast horizon is maintain as 12 Month and there are 12 historical data points. Ex-post will remain same in both the option and only difference we can see in the future forecast number as in one method it doesn't include previous generated forecast to calculate next period forecast and in other it use the previous generated forecast number to calculate next period forecast.
Extended calculation horizon checked OFF
Extended calculation horizon checked ON
Download the file by clicking below link where you see the logic for this stat model.
https://drive.google.com/file/d/1ebfDyi9BPXn0GjFXGRopxx8oJ3dfcfgl/view?usp=sharing
Weighted Average:
Details for these stat model has been maintained as shown below.
Source Weight is an input keyfigure where weightage information will be maintained to generate the appropriate forecast like demand planner could give more weightage to the most recent data that is 75% and very far off historical data were given as 25% weightage.
Output after successful completion of application job.
The formula in excel populate the same number which we can see from system generated result.
Download the logic for these stat model by clicking below link
https://drive.google.com/file/d/1Ig-1oVegWKDR-CAIcaNfNEr37ZfW5tgH/view?usp=sharing
Weighted Moving Average:
Weighted Moving Average works similar like Simple moving average algorithm where in one future forecast will remain constant and in another with the extended calculation horizon checked on, system takes historical data plus the previous generated forecast number including weightage factor.
Below is the information maintained without extended calculation horizon
Below is the information maintained with extended calculation horizon
Output after successful completion of application job.
Extended Calculation ON
Excel cell value matches with the system generated number in both the cases.
Extended calculation horizon checked OFF
Extended calculation horizon checked ON
Download the logic for these stat model by click below link
https://drive.google.com/file/d/11L9Ni0cyoDf7s8_VZ_Pm42RzcvYsvDmQ/view?usp=sharing