Model Nuggets

A model nugget is a container for a model, that is, the set of rules, formulas or equations that represent the results of your model building operations in SPSS® Modeler. The main purpose of a nugget is for scoring data to generate predictions, or to enable further analysis of the model properties. Opening a model nugget on the screen enables you to see various details about the model, such as the relative importance of the input fields in creating the model. To view the predictions, you need to attach and execute a further process or output node. See the topic Using Model Nuggets in Streams for more information.

When you successfully execute a modeling node, a corresponding model nugget is placed on the stream canvas, where it is represented by a gold, diamond-shaped icon (hence the name "nugget"). On the stream canvas, the nugget is shown with a connection (solid line) to the nearest suitable node before the modeling node, and a link (dotted line) to the modeling node itself.

The nugget is also placed in the Models palette in the upper right corner of the IBM® SPSS Modeler window. From either location, nuggets can be selected and browsed to view details of the model.

Nuggets are always placed in the Models palette when a modeling node is successfully executed. You can set a user option to control whether the nugget is additionally placed on the stream canvas.

The following topics provide information on using model nuggets in IBM SPSS Modeler. For an in-depth understanding of the algorithms used, see the IBM SPSS Modeler Algorithms Guide, available as a PDF file as part of your product download.