Model nuggets for boosting, bagging, and very large datasets
If you select Enhance model accuracy (boosting), Enhance model stability (bagging), or Create a model for very large datasets as the main objective on the modeling node, IBM® SPSS® Modeler builds an ensemble of multiple models. See the topic Models for ensembles for more information.
The resulting model nugget contains the following tabs. The Model tab provides a number of different views of the model.
| Tab | View | Description | Further Information |
|---|---|---|---|
| Model | Model Summary | Displays a summary of the ensemble quality and (except for boosted models and continuous targets) diversity, a measure of how much the predictions vary across the different models. | See the topic Model Summary (ensemble viewer) for more information. |
| Predictor Importance | Displays a chart indicating the relative importance of each predictor (input field) in estimating the model. | See the topic Predictor Importance (ensemble viewer) for more information. | |
| Predictor Frequency | Displays a chart showing the relative frequency with which each predictor is used in the set of models. | See the topic Predictor Frequency (ensemble viewer) for more information. | |
| Component Model Accuracy | Plots a chart of the predictive accuracy of each of the different models in the ensemble. | ||
| Component Model Details | Displays information on each of the different models in the ensemble. | See the topic Component Model Details (ensemble viewer) for more information. | |
| Information | Displays information about the fields, build settings, and model estimation process. | See the topic Model Nugget Summary / Information for more information. | |
| Settings | Enables you to include confidences in scoring operations. | See the topic Decision Tree/Rule Set model nugget settings for more information. | |
| Annotation | Enables you to add descriptive annotations, specify a custom name, add tooltip text and specify search keywords for the model. |