Some Statistical Applications of Centrosymmetric Matrices

Centrosymmetric matrices have been recently studied on an algebraic point of view: properties like the existence of the inverse, the expression of the determinant and the eigenspaces characterisation in the case of square matrices have been object of interest. The theoretical results obtained for this class of matrices find applications in many fields of statistics. In this study, we introduce two classes of centrosymmetric matrices that are used in probability calculus and time series analysis, namely, the transition matrices for the classification of states of periodic Markov chains and the smoothing matrices for signal extraction problems.