Model types
IBM SPSS Modeler Advantage Enterprise Marketing Management Edition supports four model types. However, Predictive models are the only type of model that you can create when you use the SPSS® Model process in an Unica Campaign flowchart. Although Unica Campaign restricts which types of models can be created in the SPSS® Model process box, any of the four model types can be used for scoring in the SPSS® Score process box.
The following table describes the four model types.
| Model type | Description | Restrictions |
|---|---|---|
| Predictive | Build a model to predict a chosen target of
interest. For example, turn a loss into a profit by targeting the best available customers for your next campaign. |
Default model type that you create from an SPSS® Model process box. The Target field must contain the text or numeric values for the model to predict. |
| Predictive Rule | Interactively build a model that identifies
segments that show a higher or lower likelihood relative to a specific
outcome. Use a predictive rule model to create rules to segment customer data. You can use the rule-finder function to automatically find segments with high or low concentrations of the TRUE value for your chosen target field. For example, look for customers who gave a positive response to your campaign, and then identify segments with a higher probability of responding. |
Cannot be created or used in an SPSS® Model process
box but can be used in an SPSS® Score process box. To build a Predictive
Rule model, you must create and build the modeling stream from the IBM SPSS Modeler Advantage Enterprise Marketing Management
Edition user interface. Supports only categorical targets. |
| Clustering | Build a model that categorizes records into
groups with similar characteristics. For example, you can segment customers based on demographic characteristics or purchasing behavior. |
To build a Clustering model, you must start IBM SPSS Modeler Advantage Enterprise Marketing Management Edition from the SPSS® Model process box. Then, change the model type to Clustering and save the modeling stream. The specified Target field from the SPSS® Model process box is used as the Evaluation field. The Evaluation field and other clustering options are described in the SPSS® Modeler Advantage Guide. |
| Association | Build a model that predicts which items are
most likely to appear together and predicts the strength of the relationship
between them. For example, identify the next most likely purchase for a customer based on the items in the current basket. |
To build an Association model, you must start IBM SPSS Modeler Advantage Enterprise Marketing Management Edition from the SPSS® Model process box. Then, change the model type to Association and save the modeling stream. The Target field from the SPSS® Model process box must specify the field that contains the content. The Content field and other association options are described in the SPSS® Modeler Advantage Guide. |
Predictive Rule models
The Predictive Rule model type is not supported in the SPSS® Model process box in an Unica Campaign flowchart. If you open IBM SPSS Modeler Advantage Enterprise Marketing Management Edition from an Unica Campaign SPSS® Model process box, you cannot create Predictive Rule models or change from another model type to a Predictive Rule model.
If you change the model type of the modeling stream to Predictive Rule, the SPSS® Model process box displays an error and will not configure. However, you can manually run Predictive Rule models through the IBM SPSS Modeler Advantage Enterprise Marketing Management Edition interface and build them interactively.
Predictive rule models can be selected for scoring in SPSS® Score process boxes.