Robust Standardizer#
This standardizer transforms continuous features by centering them around the median and scaling by the median absolute deviation (MAD) referred to as a robust or modified Z-Score. The use of robust statistics make this standardizer more immune to outliers than Z Scale Standardizer.
\[
{\displaystyle z^\prime = {x - \operatorname {median}(X) \over MAD }}
\]
Interfaces: Transformer, Stateful, Reversible, Persistable
Data Type Compatibility: Categorical, Continuous, Image, Other
Parameters#
| # | Name | Default | Type | Description |
|---|---|---|---|---|
| 1 | center | true | bool | Should we center the data at 0? |
Example#
use Rubix\ML\Transformers\RobustStandardizer;
$transformer = new RobustStandardizer(true);
Additional Methods#
Return the medians calculated by fitting the training set.
public medians() : ?array
Return the median absolute deviations calculated during fitting.
public mads() : ?array