Ansicht
Dokumentation

SAPFKCBD - Data mining processor

Fill RESBD Structure from EBP Component Structure   Fill RESBD Structure from EBP Component Structure  
This documentation is copyright by SAP AG.

Data Mining

Note

The transactions described here are not contained in the menus and can only be accessed using the transaction codes listed below.

Concept

You can use the data mining tool to analyze data in SAP-EIS and CO-PA automatically. The tool searches through a dataset for objects which stand out in some way, without the user having to carry out the usual drilldown functions manually. The tool is designed to imitate a human procedure when performing drilldown reporting. It does this by simulating drilldown in a multi-level report and subsequently issuing a list of those reporting objects considered to have striking values.

Data mining is based to a great extent on the functions contained in drilldown reporting. For this reason, the following steps which originate predominantly from drilldown reporting have to be performed:

First you need to create a data-mining form. Using this, you then create a data mining report. In addition, you must specify a data mining method which is used to control the course of the analysis path. When you start the data mining report, the system searches the dataset in the background for reporting objects which stand out in relation to the parameters you have set and subsequently lists them.

The following sections cover the objects and functions relevant to data mining.

Data mining form

In the data mining form, you specify the key figures which are to form the basis of the automatic analysis. To do this, you need to specify a search ratio and an absolute basis in the data mining form. The search ratio is the key figure to be investigated in the automatic drilldown. It is usually a relative value such as the plan/actual variance of the contribution margin or revenue. However, the absolute value of a relative key figure can vary greatly depending on the value it is based on. In order to control variations, data mining also requires you to specify an absolute key figure. You do this by choosing a key figure as an absolute base. The data mining tool then calculates the relevance figure internally based on the combination of the search ratio and absolute base. This constitutes the main ratio for analysis.

You maintain these key figures while you are maintaining the data mining form. To select the search ratio and absolute basis, position the cursor on the column containing the appropriate key figure and choose 'Search ratio' or 'Absolute basis'. The column chosen then appears in a different color.

Transactions:
- KCD5: Create data mining form
- KCD6: Change data mining form
- KCD7: Display data mining form

Data mining report

You create, change, and display data mining reports in a similar way to ordinary form reports. The only difference is that you must choose a data mining form when you create a data mining report.

However, when you execute a data mining report, you also need to choose a data mining method. When you start a report, a screen containing the parameters for the data mining method appears. By entering a name in the header, you can choose a data mining method and change it if necessary. Press the button 'Confirm' to start the data mining process which is performed internally. The system then lists the report objects it has found which stood out.

Transactions in SAP-EIS:
- KCD1: Execute data mining report
- KCD2: Create data mining report
- KCD3: Change data mining report
- KCD4: Display data mining report

Transactions in Profitability Analysis (CO-PA):
- KED1: Execute data mining report
- KED2: Create data mining report
- KED3: Change data mining report
- KED4: Display data mining report

Data mining method

The data mining method contains all the parameters required for controlling the analysis process. Creating, changing, and displaying a data mining method are all carried out on a screen in which you enter the data mining method and a second screen on which the parameter values are displayed. When you create a data mining method, you can enter a reference method and alter those entries if necessary.

A data mining method contains the following parameters and parameter groups:

Maximum runtime

The maximum runtime provokes a termination of the analysis after the amount of time specified here. If you do not enter a value, the system performs the entire analysis regardless of the time it takes. If the analysis terminates because the time ran out, the system displays a list of all the interesting report objects found until then.

Object filter

The object filter is used to limit the span of objects to be searched for during the analysis. It does this using two parameters. First you can limit the drilldown using the ranking list options by setting the type (Last n, Top n, Last %, Top %) and size of the ranking list (absolute or percentage). You also need to enter the maximum interval to the next object. On the basis of the objects sorted by relevance figure, this field determines the maximum valid difference between two neighboring objects for the next object to be investigated further. If, on the basis of the relevance figure (calculated automatically from the search ratio and absolute basis), the difference is greater than the percentage value you enter here, the list of objects relevant for data mining is cut short. The general rule is that the smaller the value entered here, the shorter the time required for analysis but the narrower the analysis.

Drilldown filter

Display filter

Weighting

On the basis of the search ratio and absolute basis, data mining calculates the relevance figure internally. This is then the main ratio for analysis. Using the weightings, you can define how strong an influence the search ratio and absolute basis are to have on the relevance figure and therefore on the analysis path.

Value filter

In this section, you can exclude certain values from the search ratio and absolute basis. This means that objects with this value are excluded from the analysis. By doing this, you can prevent the results of the analysis from being influenced by misleading values such as '0' which would cause objects to stand out for the wrong reasons.

Transactions:
- KCM1: Create data mining method
- KCM2: Display data mining method
- KCM3: Change data mining method

Results of analysis

When the analysis has been performed, the results are issued in the form of a list. This displays an overview of the striking objects found in the search. You can choose a specific line and display a detailed list containing all the key figures of the object. You can also save the results on the database. A further transaction enables you to manage the archived results. This transaction puts together a list of all the analyses saved. By double-clicking a line, you can display the overview of objects contained in that analysis (see above). For each of these, you can then display detailed information. The list consists in total of three levels. As well as displaying interesting objects which have been archived, you can also delete objects which are no longer of interest. They are then deleted from the database.

Transactions:
- TKCD8: Display (and delete) data mining results

Requirements

Output

Example

Vendor Master (General Section)   CL_GUI_FRONTEND_SERVICES - Frontend Services  
This documentation is copyright by SAP AG.

Length: 7841 Date: 20240531 Time: 145249     sap01-206 ( 184 ms )