Mean-square convergence of least-squares identification of white-noise-driven time-series models
Daniel Graupe, ELI FOGEL · International Journal of Systems Science · 1978
The paper extends previous p-lim convergence results for identification of time-series models based on a convergence-in-Lt version of Markoff's theorem to a stronger mean-aquare convergence result. The models considered may be stable or unstable, and are driven by white noise of constant variance and of bounded fourth moment, without further restriction on distribution or on the behaviour of the fourth moment.