In this paper, we focus on algorithms for Robust Procrustes Analysis that are used to rotate a solution of coordinates towards a target solution while controlling outliers. Verboon (1994) and Verboon and Heiser (1992) showed how iterative weighted least-squares can be used to solve the problem. Kiers (1997) improved upon their algorithm by using iterative majorization. In this paper, we propose a new method called “weighted majorization” that improves on the method by Kiers (1997). A simulation study shows that compared to the method by Kiers (1997), the solutions obtained by weighted majorization are in almost all cases of better quality and are obtained significantly faster.
|Title of host publication||Studies in Classification, Data Analysis, and Knowledge Organization|
|Editors||Maurizio Vichi, Paola Monari, Stefania Mignani, Angela Montanari|
|Publisher||Springer Science+Business Media|
|Number of pages||8|
|ISBN (Print)||9783319557076, 9783319557229, 9783540238096|
|Publication status||Published - 2005|
|Event||Biannual meeting of the Classification and Data Analysis Group of the Italian Statistical Society, CLADAG 2003 - Bologna, Italy|
Duration: 22 Sep 2003 → 24 Sep 2003
|Series||Studies in Classification, Data Analysis, and Knowledge Organization|
|Conference||Biannual meeting of the Classification and Data Analysis Group of the Italian Statistical Society, CLADAG 2003|
|Period||22/09/03 → 24/09/03|
Bibliographical notePublisher Copyright:
© 2005, Springer-Verlag. Heidelberg 2005.