Interaction Parameters Dialog box

When you create a linear regression, you can specify available independent variables where you suspect that the data parameters might have a combined effect on the dependent variable.

Creating interaction parameters is useful when the modeler believes the combined interaction of two independent variables is not additive.

Note: For Hadoop only, if the preceding operator is Variable Selection, then this option is disabled, because the selected columns are determined by the variable selection results.

Setting	Description
First Column	Sets the first column to be combined.
Interaction	Defines the interaction between First Column and Second Column. Can be one of the following. * - indicates the two selected variables should be defined as `a :b + a + b`. In other words, the two variables' values are multiplied together (`a:b`) and also added individually(`a, b`). For example, Gender * Adult might create a new variable with the values of `male`, `female`, `adult`, `juvenile`, `male adult`, `female adult`, `male juvenile`, and `female juvenile`. : - indicates the interaction of the two column variables are as follows. Interactions between numeric variables generate a new column that is the product of the variables. Interactions between categorical variables (word strings) generate a column per category pair, such as male/female. Interactions between numeric and categorical variables generate a new variable per category for analysis.
Second Column	Sets the second column to be combined.
New field	Adds a new row of interacting columns.
Delete	Deletes the corresponding row.