One to Many (PMML) (deprecated)
Transforms all possible values in a selected column each into a new column. The value is set as the new column's name, the cell values in that column are either 1, if that row contains this possible value, or 0 if not.
The node appends as many columns as possible values are defined for the selected column(s).
If a row contains a missing value in a selected column all corresponding new columns contain the value 0.
To avoid duplicate column names with identical possible values in different selected columns, the generated column name includes the original column name in this case (i. e. the name looks like possibleValue_originalColumnName).
The dialog of the node allows you only to select columns with nominal values. If no column name appears in the dialog but your input table contains nominal columns, you could use the DomainCalculator node and connect its output to this node.
If the optional PMML inport is connected and contains preprocessing operations in the TransformationDictionary the conversion operations of this node are appended.
- Type: Data Data
- Type: PMML Optional PMML document to add the normalization operations to.
- Type: Data Data with transformed columns
- Type: PMML PMML port object that includes the performed operations.
Manipulation > PMML
Make sure to have this extension installed:
Update site for KNIME Analytics Platform 3.7:
KNIME Analytics Platform 3.7 Update Site