Custom quantization borders and NaN modes

Contains

Custom quantization borders and the method for processing NaN values for the dataset features.

Format
  • Each line contains information regarding a single border and optionally the NaN mode settings for the corresponding feature.

  • NaN modes are output for a feature if the following conditions are met at the same time:
    • The chosen NaN mode for the feature differs from Forbidden.
    • NaN values are present in the dataset.

    The global NaN mode is specified in the --nan-mode (nan_mode) training parameter and can be overridden in the Custom quantization borders and NaN modes input file.

  • Format of a single line:
    <zero-based feature ID><\t><border value><\t><nan mode>
Example
0<\t>0.25
0<\t>0.75
2<\t>0.3<\t>Max
2<\t>0.85<\t>Max