HISTORY CLEANSING IN SAP IBP

History Cleansing in SAP IBP

In this article, we will discuss history cleansing in SAP IBP. History cleansing plays an important role in demand planning. It is one of the most important steps for a demand planner while doing demand planning.
In this article, we will discuss:

  1. Why history cleansing is required?
  2. What are the different ways/algorithms for doing history cleansing?
  3. Substitute missing values algorithm in detail.

Why history cleansing is required?

We use historical data for doing statiscal forecasting. During forecasting, it may happen that certain time series values are missing from the input data. There can be a variety of reasons for this lack of information; for example, lost records or technical issues may lead to missing data in the time series. There can be outliers in historical data. An "outlier" is a value in the historical data that lies outside the accepted range of values, which is also called the tolerance lane. There can be a variety of reasons for this deviation; for example, a data entry error or a one-time event that affected the sales results. There could be some positive outliers due to promotion sales, which also needs to be removed for short term demand planning.

These missing values, outliers need to be identified and replaced by the system so that our forecasting will not be poor and it will be having more accuracy. For doing this, SAP IBP has provided a preprocessing step while creating a forecast model whereby using different algorithms we can achieve well-prepared data for doing forecasting and improve our accuracy.

What are the different ways/algorithms for doing history cleansing?

SAP IBP has provided three different algorithms for doing history cleansing.

Interquartile Range Test

  1. Substitute Missing Values
  2. Outlier Correction
  3. Promotion Sales Lift Elimination

Substitute Missing Values

We will see different steps and settings when adding the substitute missing values algorithm to your forecast model:
Manage Forecast Model App in SAP IBP allows us to create and edit the models which are required to do forecasts in SAP IBP for demand.

We can use this app to do the following:
  1. Make general setting.
  2. Add preprocessing algorithms.
  3. Add forecasting algorithms.
  4. Add post-processing steps.
We will see these steps one by one:

Make general settings:

We maintain the following general settings in the General tab.

Model Name and Description
Model Name:
Model Description:
Time Settings

Periodicity:

Periodicity is nothing but the time profile level used by the forecast model when calculating forecast demand. If you want to forecast demand on a weekly time profile level, then you will have to choose week as periodicity. It could be a day, week, month, quarter etc.

Historical Periods:

The number of time-periods in the past which you are going to use to calculate a future forecast. Here in the image, we have taken 52 weeks as historical periods. We are fetching the last 52 weeks’ data to forecast demand.

Forecast Periods:

The number of time-periods in the future for which you want the system to calculate the forecast. Here we have taken it as 26 which means by using the last 52 weeks’ data we are calculating forecast demand for the next 26 weeks.


Add Preprocessing Algorithms

We can add any of the below algorithms as per our requirement or all at a time. For our study purpose, we will select Substitute Missing values.

You can make the following settings when adding the substitute missing values algorithm to your forecast model:

Methods used for substitution:

You can choose to substitute missing values in the selected time scope with any of the following:
  1. The mean of all other values in the historical periods you specified for the forecast model.
  2. The median of all other values in the historical periods you specified for the forecast model.
  3. A given value; that is, the number you specify in the Value for Substitution field. Here in our example, we have selected 10 as a value that will be used to replace any missing value..




Time Scope of Substitution:

By using this setting we can restrict the time frame for substituting missing values to just the past or just the future.

Key Figures:

By using this setting we can define which key figure values the system should examine to find missing values and the key figure in which the system should save the prepared data. In the Input for Algorithm We have defined Actuals Quantity as input and in Save the Result we have defined Actuals Quantity Adjusted key figure. So the system will find missing values in the actual quantity key figure and replace them with the mean of all key figure values in the past, then it will save it in the actuals quantity adjusted key figure.

Here in this article, we have discussed substitute missing values algorithm only. In the upcoming article, I will cover outlier correction and promotion sales lift elimination algorithms in detail.

SAP Integrated Business Planning for Supply Chain. (SAP Help Portal)
https://help.sap.com/viewer/feae3cea3cc549aaa9d9de7d363a83e6/2108/en-US/603f8154c2652357e10000000a44176d.html

Amol Khomne
SAP IBP Consultant, Baranwal Consultancy and Services (BCS), Pune, Maharashtra, India

Go to top

Sidebar Menu