Forecasting Models in Demantra

Hi All,
Does any one tried with only enabling certain models in Demantra. In business modeler, I enabled only regression in model library and ran the engine. I compared with demantra forecast with all the models checked. There is no change in demantra forecast. Is it the expected behaviour. Inspite of user input, demantra goes only by the recommendation from baysean model?
Thanks
Ramkri

Hi
Plz check the following
Demantra - Reciveing flat line forecast (Naive)
Forecast Engine forecasting 2.5 to 3 times more than expected
Statistical Forecast - generates the same forecast qty for all periods
Demantra: Very good historical fit -> Bad forecast
forecast tuning in demantra
Hope it will help
Tks
MJ

Similar Messages

Calculation of Safety Stock and Reorder Point under Forecast Model T

Hi Gurus!
Happy Holiday!
I would like to ask for your assistance on how the the safety stock and reorder point was calculated with the following values available. I would really appreciate it if you could give me details on the calculation.
Below are the values:
Basic data
Forecast date        01.12.2009        Unit                  CTN
Forecast model       T                 Service level         0.0
Period indicator     M                 Paramtr profile
Control data
Initialization                                Tracking limit        4.000
Model selection      A                 Procedure selection   2
Parameter optimizatio                Optimization level    F
Alpha factor         0,10               Beta factor           0,10
Gamma factor       0,00              Delta factor          0,30
Basic value           5.464-           Trend value          5.603-
MAD                      4.758            Error total              4.722
Safety stock         1                   Reorder pnt.          1
No. of values
Consumption           6                Forecast periods       1
Initial. periods      0                Fixed periods          0
Periods per season   12
Historical data
Period                Original     Corrected value
11.2009                3.000              3.000
10.2009                0.000              0.000
09.2009                0.000              0.000
08.2009                9.000              9.000
07.2009               21.000             21.000
06.2009               20.300             20.300
Forecast results
Period                Original     Corrected value
12.2009                0.000              0.000
Appreciate your assistance!
Thank you and Happy Holidays!
Ji

Sweth, you are asking for consulting, and in my opinion it is way beyond what can or should be reasonable to achieve in such a forum. You are asking complex questions, that most probable have more than one possible answer.
I suggest that you get on-site help from a knowledgeable and experienced consultant. These are crucial business issues, and should be dealt seriously.

Several forecast models

Hi,
I need to see the result of several forecast models. how to do it ?
Do I need to create n number of forecast profie to see the result of n number of forecast models or in one profile I can see the result of n number of forecast models.
Thanks a lot in advance,
Prabhat

Hi Senthil,
Thanks for your answer. I am getting lot of help from you.
You are correct that in the second tab we can change the models and see the different result based on the model we select.
But the thing is - there we have only few models to select for but when we create the forecast profile we get lot of options to select the forecast model.
Is it possible to to show the result of various forecast models under one forecast profile.or Do I need create multiple forecast profile for multiple models.
Thanks once again for your support.
Thanks,
Prabhat

Finding the forecasting model selected by Model 56

Hi,
If I use statistical model 56 and run statistical forecasting, how do I know strategy 56 is finally selecting? I see the errors and the parameters when I click the "forecast comparison" icon. But I wan to know which model the system finally selected? Is there any tale that tells me that?
How do I do the assignment of forecasting model to a selection id and forecast and see the results in the background? ANy tables, programs???
Thanks.

Hi Sinivas,
Thanks for the reply. But I am asking is to see what model 56 finally decides to go with? I mean trend or seasonal or....how can I know that the automatic model has selected a particular model after testing for all the models.
Hope this is clear.

Best way to know correct forecast model - process chain set up with multiple forecast models

Hi Experts,
I need your help in selecting best forecast model for our company. We have some of the models already used for our company, and because of multiple models used it is taking very long time for process chain to finish. There is no existing documentation available on which model was used why initially. Please help me to make out forecasting process smooth.
- What is the best way to know, which forecast model is correct and should be used for our forecasting process.
- In case multiple forecasting models are really required to be used, please suggest ways to optimally schedule them in process chain.
- At times we get messages like "not enough data available" for specific model - any way to avoid this.
- How to optimally use parallel processing profiles forecasting process in process chain.
- Things which should be avoided.
Request your help, please share your experiences.
Regards
NB

Hi Neelesh,
There are many points you need to consider to redesign forecast process for your company/client.
You need to select the best suited forecast model first depending on the business. This has to be well tested & agreed by business users. Complexity will be an outcome of this exercise with business users. Best id to give then a brief intro on all available models & then help them selection the best one as per their requirement.
Auto selection models are generally more time taking & should be used only when you have no idea at all on the business/demand pattern.
Run time will depend how you are clubbing the CVCs to get the forecast generated & also parallel processing. For parallel processing profile you will need to do trial & error testing along with help from Basis team on how many free dial up processes are available.
Even you can run many forecast calculations in parallel if the product/cvcs are totally different. - As per my personal experience maximum run time reduction can be achieved here.
Daily run is not advisable except only for businesses where you have too much dynamism in demand planning i.e. you expect the demands to be changed overnight. Most of the companies run forecast on monthly basis or at weekly basis at the max.
"Not Enough data" will be a problem if you are having the irrelevant models used in forecast profiles. This means users are not bothered to maintain the needed data for he forecast calculations or they are not aware at all of the situation. Running such models on daily basis is not advised at all. Better users should use interactive forecasting & saving the results in such cases.
Just to give a crude example we get forecast calculated on monthly basis for approximately 4 lac cvcs in less than 3 hrs using moving avg, seasonal linear regression, seasonal trend, croston models. We use parallel profiles also everywhere with 10 blocks & 500 cvc/block.
Hope this helps. Let me know if you have nay more questions & also the results using any of this.
Regards,
Rahul

Forecast model K (Constant with smoothing factor adjustment)

Hi All,
i am using forecast model K in materail master. I would like to know the system calculation on forecast value?
For Example:
Consumption qty maintained in material master for last 3 months.
Period          Total consumption
04.2011     100
03.2011     150
02.2011     200
When i am using Forecast model K, system is calculating as follows,
Period       Orig. HV   Corr. HV   Ex-post FV     Orig. FV
M 02.2011          200        200
M 03.2011          150        150        200
M 04.2011          100        100        155
M 05.2011                                                           106
05.2011     Forecast value is 106, Basic value is 105.500, Error total is -105 and MAD is 27.
I would like to know the system calculation of these value. Particularly Basic value and MAD.
And also the difference between Forcast Model D and K if possible.
When i am using Forecast model D, system is calculating as follows,
Period       Orig. HV   Corr. HV   Ex-post FV Orig. FV
M 02.2011          200        200
M 03.2011          150        150        200
M 04.2011          100        100        190
M 05.2011                                           172
Basic value 172.000,         MAD 38,         Error total -140
Thanks in advance,
Babu

MODEL K:
Constant Model with First-Order Exponential Smoothing
The constant model with first-order exponential smoothing is derived as in formula (5). A simple transformation gives the basic formula for exponential smoothing as shown in (6).
To determine the basic value, you only require the basic value from the preceding period, the last past consumption value and the alpha smoothing factor. The smoothing factor weights the most recent consumption values more than the less recent ones, so that they have a stronger influence on the forecast.
The forecast value is the basic value for the last period for which historical data is avaialble, that is the last ex-post period.
where k> n
How quickly the forecast reacts to a change in consumption pattern depends on what value you give the smoothing factor. If you set alpha to be 0, the new average is equal to the old one and the basic value calculated previously remains; that is, the forecast does not react to current consumption data. If you give alpha the value 1, the new average equals the last consumption value.
The most common values for alpha lie between 0.1 and 0.5. An alpha value of 0.5 weights past consumption values as follows:
1st historical value : 50%
2nd historical value : 25%
3rd historical value : 12.5%
4th historical value : 6.25%
and so on.
The weightings of past consumption data can be changed by one single parameter. Therefore, it is relatively easy to respond to changes in the time series.
The constant model of first-order exponential smoothing derived above is applicable to time series that do not have trend-like patterns or seasonal-like variations.
MODEL D:
I think that simple constant model is just a normal average.

Forecasting models and errors

Hi Expeing
I have following queries with regard to the forecasting Techniques:
1)How can we identify the periods in Seasonality?How to judge the seasonal periods to be taken for running the seasonal forecasting models?
2)How the errors are used in judging the forecast?As we have 6 errors what is the importance of the errors(MAD,MAPE,MSC,MPE,RMSC)?
What error should be taken as consideration while forecasting?
Why some planners only consider MAD and some planners only consider MAPE ,MSC while judging the forecast models?What is the difference between MAD,MAPE,MSC,MPE &RMSE
Please throw some light on the above questions.Your reply is highly awarded.
with regards
sai

Dear Sai,
No answer can be exhaustive .I have tried to explain as short as possible.
For Identification of periods of seasonality ,it is important to clean the historical data from abnormal high and low sales.Abnormal high sales can be caused by sporadic opportunities and low sales may happen due to natural calamities. These unnatural sales should be removed in consultation with business people to get clear picture.
Seasonal variation is measured in terms of an index, called seasonal index. It is an average that indicates the percentage of an actual observation relative to what it would be if no seasonal variation in a particular period is present
Measuring Forecast Errors
Method and Purpose
Mean squared error (MSE): Measures the dispersion of forecast errors; large errors get more weight than when using MAD. Therefore, it is sensitive to non-normal data contaminated by outliers, and such data are common.
Mean absolute deviation(MAD):Measures the dispersion of forecast errors.Measures the size of errors in units. Though it is a good measure for single product forecast, but if we aggregate MAD over multiple items, there is a possibility of high volume products dominating the results. MAD is a linear metric for error and gives same weight to all errors , large or small.
Mean absolute percent error (MAPE):Measures the dispersion of forecast errors relative to the level of demand.It measures the size of error in percentage terms. A MAPE of .19 suggests that on average the difference between the forecasted and actual values is 19%. MAPE is scale sensitive and therefore meaningless for low volume data or data with zero demand periods.
RMSE:The quadratic error provides estimates that are more linked to variance and standard deviation of demand distribution. RMSE is a quadratic metric for error and tends to overweight large errors.
Regards,
Samir Baruah

Forecast Model in Demand planning and E book Sap trainingcourse

Dear Experts ,
I am Exploring differnt Forecast Model , like constant
model, trend model, seasonal model, trend and seasonal model, Croston method
with exponential smoothing, linear regression, and causal models with multiple
linear regression (MLR).
is there any SAP links or material or course availabe to learn more about Forecasting Techniques that we can use in Demand Planning in APO
if you have idea please help me to learn .. i am really starving for this techniques
Thanks in advance
Cheers & Thanks
Raj

Hi Raj
Below link will give you entire information on forecasting models, forecasting accuracy technequues etc,:
http://help.sap.com/saphelp_apo/helpdata/en/33/437a37b70cf273e10000009b38f8cf/frameset.htm
For MLR refer the following link:
http://help.sap.com/saphelp_apo/helpdata/en/ac/216ba4337b11d398290000e8a49608/frameset.htm
Which is the best model to use:
You should try running forecasting using various model(seasonal, seasonal linear, linear, constant etc) and comapre
the forecast accuracy and find out which one is the best suitable.
Another way is to choose auto model which will select the best possible forecast strategy and give the result.
I hope thos will help you.
Please let us know if you require any more information.
Thanks
Amol

Forecast model changing

Hi
For a material in material master iam maintaining forecast model as "J".
The other parameters maintained are:
selection procedure-2
tracking limit-4
reset automatically- is checked
model selection-A
ALPHA FACTOR AND GAMMA FACTOR MAINTAINED
initialization-X
Now when i run the program RMPROG00, then in isee that the forecast model changes to "blank" or "D", "S". pls tell me how this change is happening?

Hello Dinakar,
I am also facing the similar issue, Are you able to get rid of this issue.
regards,
JPS

Automatic forecast model determination

Hi everyone, I would appreciate a lot if you can help with the following problem: in order to apply the automatic determination of the forecast model in the materials master, what other requirements are neccesary, besides that the forecast model 'J' is placed in the forecast view?, since I've been carrying out some tests of this type of automatic forecast, and apparently it only results to detect trend and constants models, but not seasonal, although, for this last type of model, also I have dealed with fields in the forecast view of material master, such as Historial Periods, Periods by Seadon, Forecast Periods, and Model Selection. When I have an historical one of consumptions with seasonal flow, if I do it with automatic forecast model, it does not detect no seasonal flow, reason why it considers a constant model. However, if I place manually that the model is seasonal, the forecast works in a correct form.
There is something that I'm avoiding in the configuration or in the updating of the materials master?
From already thank you very much.
Regards.
Luis Carbajal

Hi,
Automatic model selection procedure 1 is used in forecast strategies 50, 51, 52, 53, 54 and 55.
Automatic Model selection procedure 2 is used in forecast strategy 56.
Procedure 1 executes Seasonal and Trend tests and if neitehr of these are postive then constant model is checked. For Procedure 1 , is to recommended to be used without outlier control.
Procedure 2 executes using Constant, trend, seasonal and seasonal trend model. For procedure 2, we must remember that when you use the outlier correction, the results are not comparable with the results of the individual processes, since another procedure can be selected for the outlier correction than for the final forecast.
Linear regression and seasonal linear regression are added to automatic model selection 2 . In addition , a trend test, a seasonal test and a test of white nosie are introduced in model selection 2.
*Further the way series of tests are carried out is different for the two procedures and this is explained in SAP Help .
*http://help.sap.com/saphelp_scm70/helpdata/en/19/98ad1765354d7ba54b1eb164c377e0/frameset.htm
Hope this gives you insight.
Regards
Datta

Forecast model for Summer / Winter

Experts-
Need your help in customizing Forecast model in material master (Forecasting tab). Where these are defined in configuration? How to create new forecast model for Summer and Winter?
Here we are dealing with seasonal demands for purchasing different materials.
I am trying to figure out how to define Season codes for summer and winter? So I can flag material for Summer/Winter? As well as run forecast based on these codes? I understand that I can define season model or trend model, but in that case I will be unable to define which material is for summer and which one is for winter?
Looking forward for your inputs-
Thanks in advance.

Started using MRP controller for summar and winter parts.

Forecasting Models in Demand Planning.

Hi experts,
want to know in details and also to practice various forecasting models in Demand Planning, from beginning.
please suggest.
Regards
Bhupendra

Hi Bhupendra,
There are various forecasting models to use in business like automatic model and manual model as a broad category.
In detailed, there are models like constant model, trend model, season models, trend seasonal etc.,
Forecasting models are very simple and mostly based on STATISTICAL TECHNIQUES. If you have sound knowledge in various statistical methodologies, it will be very easy to followup all the models.
Moreover, while selecting the best model for an organisation, one needs to study and identify the key performance indicators for improvement and accordingly the forecast model needs to be aligned.
The method to find this kind of model is the BEST FIT METHOD (it could be one or more or combo of it).
In demo systems or sandbox system, you can populate various data with various models and try to analyse the one which satisfies business requirements. You can also take the help of BUSINESS PLANNING ANALYSER to build the customised forecasting model which can be part of business side or from consulting side.
Regards
R. Senthil Mareeswaran.

Need help with seasonal forecast model

I am really struggling to get accurate forecast results for CBP using the seasonal forecast model. Our business has relatively constant consumption for 9 months of the year, say an average of 200,000 quantity per month. Then as we enter the 'holiday the season', there is a rapid increase in consumption to as much as 1.6 million per month in December.
Seasonal model seems to work fine if my period indicator is monthly. But our consumption period and forecast needs to be daily. We have enough history going back to January 2010. Can somebody suggest what forecast view settings I should be trying? ... historical periods, intialization periods, periods per season, etc??
I just can't seem to get the same forecasted holiday spikes for 2011 holiday season (Oct - Dec) or 2012.
Thanks, Vlad

Hi B.
to expose an Adaptive RFC model to other DCs you must explicitly add the following entities to the public part of your DC:
- Adaptive RFC Model entitiy
- Logical Dictionary
The Public Part Purpose must be defined as "Compilation".
Important: To expose the Public Part to another DC you must explicitly trigger a DC Build (context menu item 'Development Component -> Build ...', so that the Public Part Archives is being generated. Without an existing PP archive the Client DC cannot 'see' the exposed Public Part Entities of the Service DC.
Regards, Bertram

DP - Forecasting Models

hi
I need to work on forecasting models in my project. I read a lot of times on this subject various articles. I am still not understanding clearly the terms like Basic, trend, initialisation, Ex-forecast.. Is Ex Forecast generated by system automatically ? if so, what we should interpret from that ? its the forecast run in the past for comparison purposes. So if i run a forecast last month (oct 07) becomes ex forecast in Nov 07 ? What & How to do the Initialization ?
Hope someone can pass the light on this
thanks
venkatesh

Hi Venkat,
The basic idea behind using a statistical model is to analyze the historical sales trend and find a analogous mathematical equation there y use statistics to solve the equation. This equation depending on how the trend is have few constants. Like an equation y=ax1bx2cx3+..... where y is your forecast and x1, x2 and so on are independent variables and a,b,c,d are all constants.
The statistical models here are using apha, beta, gamma and permo in place of these a,b,c,d. These indexes are used based on the curve.
alpha index for basic value
Beta index for Trend
Gamma index for seasonal
persmo index for outliers
Any curve always has basic value as it has to define where it starts from the base. If the curve is heading up or down, it has a trend (upward or downward) value. If the curve has a seasonal behavior like a sine or cosine curve, it also has a seasonal index associated with it.
Basic: This value decides the vertical placement of the curve with respect to base. Higher the alpha(basic), higher the curve with respect to horizontal axis.
Trend: Specifies the slope of the line
Seasonal: Specifies divergence from basic value.
Ex-post forecast: In the history available, you date back some time and consider part of the history as future. Then the history before this so considered forecast is used to forecast the future. You get a future curve and value. But you already have the future as it is history. This helps you compare the effectiveness of the forecasting model.
Read this
<a href="https://www.sdn.sap.comhttp://www.sdn.sap.comhttp://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/docs/library/uuid/7a4025f8-0a01-0010-0fb2-b7ab22597675">document</a>
This is the best source.
If you select the ex-post forecast in univariate forecast profile, the system calculates the expost forecast and can be displayed below the history to compare.
Hope this helps.

APODP How Forecast model for different product can be define ?

Hi Gurus,
I had forecast 1000 products and in that thousand products i have 300 slow moving products and 700 fast moving products.
I had taken Linear regression model for this products and run the forecast. My question is i wanted to forecast 300 slow moving products with Seasonal Model , can anyone please mention how we can forecast 1000 products with two different forecast models?
Waiting for your answers
Regards & Thanks
Raj

Hi Raj,
A little concept clarification here. The was forecasting model works is"
1. You create a forecasting model giving a key figure for which you want to forecast and assign a univariae or MLR or composite profile to it assigning a key figure which you want to use as history and a statistical model. (TCODE: MC96B)
2. Create a selection profile in interactive planning screen (TCODE: SDP94). Select all the products you want to plan for in this profile along with other selections.
3. Assign forecasting profile to the profile. This way you can assign different forecasting models to different profiles.
In your case,
1.Create two forecasting profiles one with Linear regression and the other with Seasonal model.
2. Create 2 selection profiles one with the 300 products and other with the other 700 products.
3. Assign Linear regression to one selection profile and seasonal model to another profile.
Hope this elaborates the solution. Please let us know if you have any other questions.
Thanks.

Forecasting Models in Demantra

Similar Messages

Maybe you are looking for