Unreleased

OutSeekR 1.1.0 - 2026-01-15

Major Changes

• Breaking change: Replaced GAMLSS-based distribution fitting with Gaussian Mixture Models (GMM) using the mclust package for more flexible and robust distribution modeling.
• Breaking change: Removed distributions output from detect.outliers() function as individual distribution fitting is no longer performed.
• Breaking change: Removed distribution parameters from calculate.residuals(), simulate.null(), and outlier.detection.cosine() functions.

Improvements

• Enhanced null data simulation using GMM-based approach that better captures complex data distributions.

Dependencies

• Added mclust package dependency for Gaussian mixture modeling.
• Removed dependencies on gamlss and gamlss.dist packages.

Documentation

• Updated methodology descriptions in README and vignettes to reflect GMM-based approach.
• Revised examples to remove distribution-specific parameters and outputs.

Function Changes

• detect.outliers(): Removed distributions from output list.
• calculate.residuals(): Removed distribution parameter, now uses GMM fitting.
• simulate.null(): Removed x.distribution and r.distribution parameters, now uses GMM sampling.
• outlier.detection.cosine(): Removed distribution parameter, now uses GMM-based quantiles.
• calculate.p.values(): Removed x.distribution parameter.