On the AR modelling of the one-sided autocorrelation sequence for noisy speech recognition | Publicación