Skip to content

SBT Feature Transformer

A feature engineering module that encodes sample using leaf indices predicted by Hetero SBT/Fast-SBT. Samples will be transformed into sparse 0-1 vectors after encoding. See original paper for its details.

Figure 5: Encoding using leaf indices\</div>

Param

sbt_feature_transformer_param

Classes

SBTTransformerParam (BaseParam)
Source code in federatedml/param/sbt_feature_transformer_param.py 
class SBTTransformerParam(BaseParam):

    def __init__(self, dense_format=True):

        """
        Parameters
        ----------
        dense_format: bool
            return data in dense vec if True, otherwise return in sparse vec
        """
        super(SBTTransformerParam, self).__init__()
        self.dense_format = dense_format

    def check(self):
        self.check_boolean(self.dense_format, 'SBTTransformer')
Methods
__init__(self, dense_format=True) special

Parameters:

Name Type Description Default
dense_format bool

return data in dense vec if True, otherwise return in sparse vec

 True
Source code in federatedml/param/sbt_feature_transformer_param.py 
def __init__(self, dense_format=True):

    """
    Parameters
    ----------
    dense_format: bool
        return data in dense vec if True, otherwise return in sparse vec
    """
    super(SBTTransformerParam, self).__init__()
    self.dense_format = dense_format
check(self)
Source code in federatedml/param/sbt_feature_transformer_param.py 
def check(self):
    self.check_boolean(self.dense_format, 'SBTTransformer')
Last update: 2021-11-08
Back to top