IBM Data WH PCA Model Nuggets

The PCA model nugget provides a means of setting options for scoring the model.

When you run a stream containing a PCA model nugget, by default the node adds one new field, the name of which is derived from the target name.

Table 1. Model-scoring field for PCA
Name of Added Field	Meaning
$F-target_name	Predicted value for current record.

If you specify a value greater than 1 in the Number of principal components ... field on either the modeling node or the model nugget and run the stream, the node adds a new field for each component. In this case the field names are suffixed by -n, where n is the number of the component. For example, if your model is named pca and contains three components, the new fields would be named $F-pca-1, $F-pca-2, and $F-pca-3.

You can view the extra fields by attaching a Table node to the model nugget and running the Table node.

Note: An error may occur when scoring the model if lowercase field names are used. This is a known Db2 Data Warehouse defect, with the workaround being to rename all the fields to uppercase before scoring.