Build a Predictive Model

Contents

• Introduction
• Steps to create a Predictive Scenario

Introduction

Create a Predictive Scenario

SAP Analytics Cloud Predictive Forecasting helps us understand past data trends to predict future metrics. The predictive algorithm classifies existing information, identifies outliers, and brings to the surface relationships within your data, to help us see and understand our business’ key influencers.

Business Goals

CFO wants to understand the financial position of the company by analyzing overdue receivables and future due receivables. SAP Analytics Cloud’s Predictive scenario capability based on regression model helps to gain further insights whether the invoice will be paid on time or late. This predictive insight would help CFO’s plan cash flow accordingly.

Technical Goals

We would be learning how to create a Predictive Scenario in SAP Analytics Cloud.

Steps to create a Predictive Scenario

1. Click on Main Menu >> Predictive Scenario.

2. Choose Regression.

3. Save new Predictive Scenario. In the Name text box type “PredictiveScenario_{userid}” with your username as suffix.

4. Click on select Training data source.

5. In the search box type CASHFLOW_AR_CLEARED_DATA.

6. Select CASHFLOW_AR_CLEARED_DATA data source.

7. Select the choose icon for Target.

8. Select “LATE_DAYS”.

9. Select the choose icon for Exclude As influencer.

10. Choose ACCOUNTINGDOCUMENT, ACCOUNTINGDOCUMENTITEM, NETDUEDATE, CLEARINGDATE.

11. Click SELECT then click Train.

12. Expand the Predictive Models(1) drop down at the bottom of the screen.

13. Ensure the model status has changed to Trained.

14. Click on Apply Predictive Model at the top.

15. In the data source text box, verify CASHFLOW_AR_OPEN_DATA is selected.

16. In the Replicated Columns text box click on choose.

17. Select all columns and click Select.

18. In Statistics and Prediction choose Assigned Bin , Outlier Indicator and Predicted Value.

19. In the Output As text box click on Choose.

20. Enter name as CASHFLOW_ARDATA_W_PREDICTION_{userid} with your username as suffix.

21. Click OK.

22. Click Apply.

Once the model is applied, you will see the status in the bottom panel changing from Applying then to Applied. This means the output is now available for consumption.

Let’s browse to find the Predictive dataset that has been created.

23. Click on Main Menu >> Files.

24. In the search box type CASHFLOW_ARDATA_W_PREDICTION_{userid}.

25. Click Search.

26. Choose the dataset and browse through the dataset.

Dataset: