If you set this indicator, the default score is issued for any data records that do not fall in the area of definition for the trained function. If you do not set this indicator, then a regression is performed for these data records independently of the discrete model fields.
Dependencies
Example
A model has two discrete fields: Occupational Group und Vehicle Category. In the training data set, there are no data records for the combination Computer Programmer and Sportscar (or the amount of data records is less than the number specified in the parameter Minimum Number of Data Records). In this case, the Scoring function is not trained using this data area, and the area of definition for the trained scoring function excludes data records containing this combination of values. During the prediction, a data record with the combination Computer Programmer and Sportscar is treated as follows: If the Prediction only in trained domain indicator is set, then the default score is issued. If the indicator is not set, then a regression is performed for this data set independently of the discrete model fields. This means that you receive the same value for the data set as the value that a model without the discrete fields would deliver.