bach.DataFrame.scale

scale​

(with_mean=True, with_std=True)

​[source]

Standardizes all numeric series based on mean and population standard deviation.

Parameters​

  • with_mean (bool) – if true, each feature value will be centered before scaling
  • with_std (bool) – if true, each feature value will be scaled to unit variance

Returns​

DataFrame

Return type​

bach.dataframe.DataFrame

Each transformation per feature is performed as follows:

In:
feature.to_pandas()
Out:
index
a           1
b           2
c           3
d           4
Name: feature, dtype: int64

>>> scaled_feature = feature.copy()
>>> if with_mean:
...     scaled_feature -= mean_feature

>>> if with_std:
...     scaled_feature /= std_feature

>>> scaled_feature.to_pandas()
index
a   -1.341641
b   -0.447214
c    0.447214
d    1.341641
Name: feature, dtype: float64

Where:

  • feature is the series to be scaled
  • mean_feature is the mean of feature
  • std_feature is the (population-based) standard deviation of feature