A fast computational algorithm for the QR-like decomposition of the modified covariance method of linear prediction

Summary form only given. A fast QR algorithm for the batch modified covariance method of linear prediction that is the companion to the fast Cholesky algorithm has been developed. A Fortran program that performs the QR solution for the linear prediction parameters has also been developed. The numerical accuracies of the fast QR algorithm and fast Cholesky algorithm solutions for the least-squares modified covariance equations that been compared.<
>