Creating, Changing, and Activating a Model 

Use

You create a model for a data mining method so that you can apply the method according to your business requirements. You use model fields in a model to specify what is to be predicted and which data should form the basis of the prediction.

You can create a data mining model using the Data Mining Workbench or the Analysis Process Designer (APD). Once you have created and saved the model to meet your requirements, you can activate it.

Prerequisites

You must have been assigned to the role Customer Behavior Analysis (SAP_BW_CUSTOMER_BEHAVIOR) and you must have chosen Customer Behavior Analysis → Customer Behavior Modeling in the user menu.

Creating a Model in the Data Mining Workbench

...

       1.      Position the cursor on a data mining method and use the right-hand mouse to choose Create in the context menu.

       2.      In the step Create Model, enter a name and a description for the model. The method name for which you are creating a model is displayed. You have three options for model field selection:

·        To create the model fields manually, select the Manual option.

·        If you want to create a model that is similar to an existing model created previously, you can copy it choosing the Use Model as Template option. You can make minor changes to the copied version manually to suit your requirements.

·        To create a model from a query, choose Model Field Selection and select the query which you want use as a source for model fields.

Selecting a query at this point will assist you in creating the model. It is therefore recommended to enter the same query that you would like to use subsequently, while training the model. However, this is not essential. You can also use any other query as a template for your model.

The InfoObjects contained in the selected query are available in the next step as model fields.

       3.      In the step Select InfoObjects, select from the query those InfoObjects that you would like to use as model fields.

If you would like to use other fields from the query as calculated or restricted key figures in your model, you need to include an SAP BW dummy InfoObject as a model field for each one. You can then assign the corresponding field to this model field in the Change mode.  

       4.      In the step Edit Model Fields, specify the attributes for each field.

¡        The description you give the model field does not necessarily have to be identical with that of the InfoObject.

¡        The system automatically copies the attributes Data Type and Length from InfoObject (these cannot be modified).

¡        The value types valid for a model field are dependent on the method that you are creating the model for and on the data type of the model field.

The value type specified for a model field determines which entries can be made as Field Parameters and Field Values.

The attributes for a model field that are listed below do not apply in the data mining method Association Analysis. No prediction is involved with this method. Instead, the association rules are determined by training and form the result. Consequently, the settings for the field parameters and field values do not apply.

¡        Set the Prediction Variable indicator for the model field for which the subsequent prediction is to be made. Select as a prediction variable that model field for which you wish to gain more information (via the model).

With the data mining method Clustering, the cluster is always the prediction variable. Consequently, you cannot specify a prediction variable for this method.

¡        The field parameters are dependent on the value type of the model field and on the data mining method.

You cannot select any parameters for model fields where the value type KEY has been set.

¡        Under Field Values, you can specify how the system should interpret specific values that can be taken by a model field but have no bearing on the result.

       5.      In the Model Parameters step, enter the parameters that are valid for the entire model. The model parameters are dependent on the data mining method.

       6.      Save the model.

Result

You have performed all necessary steps for the creation of a model. The created model appears in the tree beneath the relevant method.

Changing the Model

You can make changes to the model that you have created.

...

       1.      Position the cursor on a model that you wish to change and use the right-hand mouse to choose Change in the context menu.

       2.      In the Model Fields tab page, make your changes to the model fields. You can change the attributes for the model fields or add more model fields.

       3.      In the Model Parameters tab page, make your changes to the model parameters.

       4.      Save your changes.

Activating the Model

Once a model meets your requirements, you can activate it. The active model is then used for creating other versions. This means that, when you change a model that has been activated, the active version remains unchanged and the changes are saved in a Revised version. The active version is only overwritten when you activate the modified version.

You can only train or valuate a model or use it for the prediction if the model has been activated.

If a model has a modified version, the model name in the tree is marked in blue.

To activate a model, proceed as follows:

...

       1.      Position the cursor on a model that you wish to activate and use the right-hand mouse to choose Activate in the context menu.

The version displayed under Model Information is changed to Active.

       2.      Make any necessary changes to the model and save your changes.

The version displayed under Model Information is changed to Revised.

       3.      To navigate between the active and modified versions, place the cursor in the model and choose .