denoiseR: A Package for Low Rank Matrix Estimation

Julie Josse (1, 2) , Sylvain Sardy (3) , Stefan Wager (4)

1 LMA2 - Laboratoire de Mathématiques Appliquées Agrocampus
2 IRMAR - Institut de Recherche Mathématique de Rennes
3 Section de mathématiques [Genève]
4 Department of Statistics [Stanford]

We present the R package denoiseR dedicated to low-rank matrix estimation. First, we briefly review existing methods including singular value shrinkage and a parametric bootstrap approach. Then, we discuss how to extend the methods to missing values and suggest a general iterative imputation algorithm. It includes an extension of the Stein Unbiased Risk Estimate to missing values for selecting tuning parameters. Finally, we compare and apply the methods with many experiments.

Mots clés: bootstrap; SURE; count data; correspondence analysis; clustering; missing values; matrix completion; Low-rank matrix estimation; singular values shrinkage

Domaines: Statistiques [math.ST]

Julie Josse, Sylvain Sardy, Stefan Wager. denoiseR: A Package for Low Rank Matrix Estimation. 2016. ⟨hal-01278417⟩