SLRM Model Nuggets

Note: Results are only shown on this tab if you select both Include model assessment and Display model evaluation on the Model options tab.

When you run a stream containing an SLRM model, the node estimates the accuracy of the predictions for each target field value (offer) and the importance of each predictor used.

Note: If you selected Continue training existing model on the modeling node Model tab, the information shown on the model nugget is updated each time you regenerate the model.

For models built using IBM® SPSS® Modeler 12.0 or later, the model nugget Model tab is divided into two columns:

Left column.

• View. When you have more than one offer, select the one for which you want to display results.
• Model Performance. This shows the estimated model accuracy of each offer. The test set is generated through simulation.

Right column.

• View. Select whether you want to display Association with Response or Variable Importance details.
• Association with Response. Displays the association (correlation) of each predictor with the target variable.
• Predictor Importance. Indicates the relative importance of each predictor in estimating the model. Typically you will want to focus your modeling efforts on the predictors that matter most and consider dropping or ignoring those that matter least. This chart can be interpreted in the same manner as for other models that display predictor importance, though in the case of SLRM the graph is generated through simulation by the SLRM algorithm. This is done by removing each predictor in turn from the model and seeing how this affects the model's accuracy. See the topic Predictor Importance for more information.