STARRS

MaRDI QIDQ7374075CRANFDO

Authors Antoine Godichon-Baggioni, Daphné Giorgi, Stéphane Robin

Last update 29 July 2026

Copyright license GNU General Public License, version 2.0, GNU General Public License, version 3.0

Software version identifier 1.0

Description

Algorithms for robust multivariate statistics (STochAstic Robust multivaRiate Statistics) including geometric median and geometric median covariance computation, k-medians clustering and robust median Principal Compenents Analysis (PCA), robust estimation of parameters for Gaussian, Student, or Laplace mixture models. 'STARRS' provides an independent, clean, consolidated and cohesive framework, while drawing inspiration from the approaches implemented in packages 'Gmedian', 'Kmedians', 'RGMM', 'RobRegression'. Methods used in the package refer to H. Robbins, S. Monro (1951) <doi:10.1214/aoms/1177729586>; D. Kraus, V. M. Panaretos (2012) <doi:10.1093/biomet/ass037>; H. Cardot, A. Godichon-Baggioni (2015) <doi:10.48550/arXiv.1504.02852>; A. Godichon-Baggioni, S. Robin (2024) <doi:10.1007/s11222-023-10362-9>.

Cites work

• A robust model-based clustering based on the geometric median and the Median Covariation Matrix
• A Stochastic Approximation Method
• Dispersion operators and resistant second-order functional data analysis

This page was built for software: STARRS