Canonical polyadic tensor decomposition in the presence of non Gaussian noise

In this paper we describe an estimator for the canonical polyadic (CP) tensor model using order statistics of the residuals. The estimator minimizes in an iterative and alternating fashion a dispersion function given by the weighted ranked absolute residuals. Specific choices of the weights lead to either equivalent or approximate versions of the least squares estimator, least absolute deviation estimator or least trimmed squares estimators. For different noise distributions, we present simulations comparing the performance of the pro posed algorithm with the standard least squares estimator. The simulated performance is equivalent in the Gaussian noise case and superior when the noise is distributed according to the Laplacian or Cauchy distributions.

[1]  Morten Mørup,et al.  Applications of tensor (multiway array) factorizations and decompositions in data mining , 2011, WIREs Data Mining Knowl. Discov..

[2]  Nikos D. Sidiropoulos,et al.  Robust iterative fitting of multilinear models , 2005, IEEE Transactions on Signal Processing.

[3]  Rasmus Bro,et al.  Multi-way Analysis with Applications in the Chemical Sciences , 2004 .

[4]  R. Bro,et al.  PARAFAC and missing values , 2005 .

[5]  Zbynek Koldovský,et al.  Cramér-Rao-Induced Bounds for CANDECOMP/PARAFAC Tensor Decomposition , 2012, IEEE Transactions on Signal Processing.

[6]  P. Comon,et al.  Tensor decompositions, alternating least squares and other tales , 2009 .

[7]  M. Hubert,et al.  Detecting outlying samples in a parallel factor analysis model. , 2011, Analytica chimica acta.

[8]  J. McKean,et al.  The Geometry of Robust Procedures in Linear Models , 1980 .

[9]  Ieee Staff.  2017 25th European Signal Processing Conference (EUSIPCO) , 2017 .

[10]  Mia Hubert,et al.  Robust PARAFAC for incomplete data , 2012 .

[11]  Nikos D. Sidiropoulos,et al.  Parallel factor analysis in sensor array processing , 2000, IEEE Trans. Signal Process..

[12]  Ananthram Swami,et al.  Cramer-Rao bounds for deterministic signals in additive and multiplicative noise , 1996, Signal Process..

[13]  Pierre Comon,et al.  Tensors : A brief introduction , 2014, IEEE Signal Processing Magazine.

[14]  Eric C. Chi,et al.  Making Tensor Factorizations Robust to Non-Gaussian Noise , 2010, 1010.3043.

[15]  Nikos D. Sidiropoulos,et al.  Blind PARAFAC receivers for DS-CDMA systems , 2000, IEEE Trans. Signal Process..

[16]  Nikos D. Sidiropoulos,et al.  Cramer-Rao lower bounds for low-rank decomposition of multidimensional arrays , 2001, IEEE Trans. Signal Process..

[17]  Peter J. Rousseeuw,et al.  Robust regression and outlier detection , 1987 .

[18]  H. Vincent Poor,et al.  Robust iteratively reweighted Lasso for sparse tensor factorizations , 2014, 2014 IEEE Workshop on Statistical Signal Processing (SSP).

[19]  S. Frosch Møller,et al.  Robust methods for multivariate data analysis , 2005 .

[20]  B. Ripley,et al.  Robust Statistics , 2018, Wiley Series in Probability and Statistics.

[21]  Louis A. Jaeckel.  Estimating Regression Coefficients by Minimizing the Dispersion of the Residuals , 1972 .

[22]  J. Jurecková,et al.  Nonparametric Estimate of Regression Coefficients , 1971 .

[23]  Joseph W. McKean,et al.  Rank regression with estimated scores , 1997 .

[24]  E. H. Lloyd.  LEAST-SQUARES ESTIMATION OF LOCATION AND SCALE PARAMETERS USING ORDER STATISTICS , 1952 .

[25]  P. Paatero.  Least squares formulation of robust non-negative factor analysis , 1997 .