A simplified approach to inverting the autocovariance matrix of a general ARMA(p,q) process

This article demonstrates how to compute the exact inverse of the autocovariance matrix and its determinant more efficiently than the previous work for a general ARMA(p,q) process of length n, when n[greater-or-equal, slanted]max{p,q} is considered. We formulate the results as analytic matrix expressions, which can be easily implemented in general practice.