NEWS

---

power.transform 1.0.4 (2026-03-02)

Changes

• Updated citation, since the manuscript was published in Machine Learning: Zwanenburg, A. and Löck, S. (2026) Location and scale-invariant power transformations for transforming data to normality, Machine Learning, 115(3). Available at: https://doi.org/10.1007/s10994-026-06994-3.

power.transform 1.0.3 (2026-01-15)

Changes

• Fixed an error that occurred when assessing if a transformation should be rejected, but the empirical central normality test p-value could not be computed.

power.transform 1.0.2 (2025-10-29)

Changes

• ecn.test was renamed to ecn_test. It now additionally takes tau, n, and kappa as arguments (see function documentation).

• cn.test was deprecated. Use ecn_test with kappa = 1.0 instead.

• When checking if required packages are installed, power.transform now caches results. This prevents unnecessary look-up.

power.transform 1.0.1 (2025-04-12)

Changes

• The empirical central normality test is now dependent on the number of samples. The test can now be called using ecn.test. cn.test is a stricter central normality test whose test statistics are determined from strictly normal distributions, instead of normal distributions with up to 10% outliers.

• The robust location- and shift-invariant transformations now use weights optimised for achieving central normality.

power.transform 1.0.0 (2024-09-13)

This is the initial public release of the power.transform package.