ERROR OF MEASURES IN DP

[Former Member]

helo gurus

IN demand planning these below are the Error Measure.. These are we maintain the accuracy of Univerate forecast profile??

Mean absolute deviation (MAD)

Error total (ET)

Mean absolute percentage error (MAPE)

Mean square error (MSE)

Square root of the mean squared error (RMSE)

Mean percentage error (MPE)

My question is When we maintain these Error measure? in which scenerio???

Which is the best one to Univerate forecast profile??

Do we use more than one (above given) in Univerate Forecast profile???

In which situation do we use MAD????

plz help me out .... its gr8

[Former Member · Accepted solution]

Hi Dallyanusha,

The selection of forecast error scenario depends on the

KPI (key performance indicator) that the business wanted to

adopt and measure it. There is no specific criteria for selection

and depends purely on business

We can use one or more depends upon the measurement

that business requires it

For MAD, if you want to measure the difference between

forecasted values and historical values, then you can use

MAD selection to identify the difference.

Regards

R. Senthil Mareeswaran.

[Former Member · Accepted solution]

What are the advantages and disadvantages of using a Univariant model that leverages MAD, MAPE or both MAD & MAPE as model parameters?

[Former Member · Accepted solution]

Hi Shaughe,

Apart form the infomration given earlier, following information may be useful to you:

MAD : Measures average absolute deviation of forecast from actuals

1 Measures absolute error. It can be positive or negative

2 We want MAD to be as small as possible

3 Since this is a number, there is no way to know if MAD error is large or small in relation

to the actual data

MAPE: Measures absolute error as a percentage of the sales.

Same as MAD exceptu2026

1Measures absolute deviation of forecast from actual as a percentage of actual

data

2 Indicates persistent absolute error in forecast

3 Combinations with very small or zero volumes can cause large skew in results

4. Most common measure of forecast accuracy

Recommendation: Use percentage error calculation

1. Calculations can be easily aggregated to variety of levels

2. Initially evaluate error at high level and dig down as necessary

There may not be any added advantage to use both the errors measurements simultaneously, as one is sufficient.

Hope this helps.

Regards

Datta Kadam

[Former Member]

hi dallyanusha,

Though the question is answered , I would highlight on the use of the error parameters.

We use these error parameters to measure the accuracy of the choosen strategy in the univariate profile.

the different parameters are already explained in the replies.

By changing the forecasting parameters you run the forecast several times and then check the run which results the least error corresponding to your accuracy parameter(MAD, or MAPE etc), we finalize that strategy and paramter values(alpha/beta/gamma) to be the best suitable for our forecast run in business.

Hope this adds some value on the discussion.

Thanks

Binod