Normalization

Normalization consists of transforming numeric values into a specific range, such as [–1.0,1.0] or [0.0,1.0] so that x_new = (x_old-shift)/scale. Normalization usually results in values whose absolute value is less than or equal to 1.0.

To normalize a column:

1. In the Transform Type field, select the option Normalization.

2. In the Normalization Type field, select a type from the drop-down list. Oracle Data Miner supports these types of normalization:

   • Min Max: Normalizes the column using the transformation x_new = (x_old-min)/ (max-min). The default is min-max.

   • Z-score: Normalizes numeric columns using the mean and standard deviation computed from the data. Normalizes each column using the transformation x_new = (x-mean)/standard deviation.

   • Linear Scale: Normalizes each column using the transformation x_new = (x-0)/ max(abs(max), abs(min)).

   • Manual: Defines normalization by specifying the shift and scale for the transformation x_new = (x_old-shift)/scale. If you select Manual, then specify the following:

     • Shift

     • Scale

3. After you are done, click OK.