TwoStep-AS Cluster Model Nuggets

The TwoStep-AS model nugget displays details of the model in the Model tab of the Output Viewer. For more information on using the viewer, see Working with output.

The TwoStep-AS cluster model nuggets contain all of the information captured by the clustering model, as well as information about the training data and the estimation process.

When you run a stream containing a TwoStep-AS cluster model nugget, the node adds a new field containing the cluster membership for that record. The new field name is derived from the model name, prefixed by $AS-. For example, if your model is named TwoStep, the new field will be named $AS-TwoStep.

A powerful technique for gaining insight into the TwoStep-AS model is to use rule induction to discover the characteristics that distinguish the clusters found by the model.

For general information on using the model browser, see Browsing model nuggets.

Note: In the TwoStep-AS model viewer, under the Evaluation > Model Quality section, the number of records is displayed for each cluster. If you connect a Distribution node to calculate the score result, you may find that the number of records in each cluster is different from what you see under Model Quality section. This functions as it should. From the algorithm perspective, the evaluation result comes from a hierarchical clustering process, while the scoring table comes from directly comparing the data case with the distribution of the final clusters. These two different scoring processes may result in different results if the clustering model is not perfect. But in most cases, the difference is quite small.