What question did this study set out to answer?

To develop robust location- and scale-invariant power transformations for stabilizing variance and achieving normality.

February 19, 2026Machine LearningOpen Access

Location and Scale-Invariant Power Transformations for Transforming Data to Normality

View Full Paper

Ask AI

Bookmark

Authors

Analytical methods improve normality transformations in varied datasets, suggesting potential applications in data processing.

To develop robust location- and scale-invariant power transformations for stabilizing variance and achieving normality.
Presented location- and scale-invariant Box-Cox and Yeo-Johnson transformations.
Derived maximum likelihood estimation for optimizing transformation parameters.
Proposed robust adaptations to minimize outlier influence.
Conducted empirical testing for assessing central normality of features.
Robust transformations showed improved performance in achieving central normality in simulations.
In a machine learning experiment, the new transformations did not significantly enhance or hinder model performance compared to conventional methods.

Zwanenburg et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark