Baseline

• List the baseline value for each object on a new line.
• Use a tab as the delimiter to separate the columns on a line.
• Use the RawFormulaVal prediction type of the model values as the baseline. Other prediction types are not supported.

• The first line contains the description of data in the corresponding column.

  Format:

  • Non multiclassification methods
  • Multiclassification methods

  RawFormulaVal[<\t><other tab-separated columns>]

  RawFormulaVal:Class=<class_name_1><\t>...<\t>RawFormulaVal:Class=<class_name_N>[<\t><other tab-separated columns>]

  class_name_1...class_name_N are the names of corresponding classes. The order must match the one specified in the --classes-count parameter and in the model.

  other tab-separated columns can be set for auxiliary data (for example, for SampleId).

• All the other lines contain baseline values for the objects of the input dataset. Specify the baseline value for the i-th object on the line numbered i+1.

  Format:

  • Non multiclassification methods
  • Multiclassification methods

  <value>[<\t><other tab-separated data>]

  <value for class_name_1><\t>...<\t><value for class_name_N>[<\t><other tab-separated data]

  class_name_1...class_name_N are the names of corresponding classes.

RawFormulaVal
-2.036136239767074585e-01
-2.093112945556640625e+00
-1.437631964683532715e+00
9.149890393018722534e-02