Unit
Classifier

Declaration

TDiscreteRecord = packed record
      Values: TIntegerArray;
      ValuesCount: integer;
      ExamplesCount: integer;
      MaxValueIndex: integer;
      Weight: double;
      Enabled: boolean;
    end;

Description
Stores the parameteres learned about the specific discrete attribute during the learn phase. This includes number of samples encountered. Total examples count, and of course, the statistics about the occuring values.

Values

Values	Each index of Values array containes the number of occurences of the discrete value of the attribute, associated with this index.
ValuesCount	Total number of different values stored by the Values array. (=Length(Values)
ExamplesCount	Number of different learn examples included. This does not count the missing values.
MaxValueIndex	The index pointing to the Values array field with the highest value. This index therefore also points to the value of the discrete attribute, which occured most often.
Enabled	True, if the attribute is to be included in the classification process.
Weight	This value is 1 by default. It may be used to emphisize the value of the specific attribute towards the others.