Robust estimator of the correlation matrix with sparse Kronecker structure for a high-dimensional matrix-variate (Q2306279)

The estimation of the correlation matrix of a higher-dimensional matrix-variate \(X \in \mathbb{R}^{p\times q}\) is needed in practice many times. The authors propose a robust estimator based on Kendall's correlation. The proposed estimator is extended further to tensor data. In this paper, it is shown that the Kronecker structure actually increases the effective sample size and leads to a fast convergence rate. They apply the method to Aries of GeneExpression in the Monse Aging (AGEMAP) database to investigate the behaviour of the proposed method.