I'm working on a question regarding the convergence range of the spectrum of a discrete-time stationary process, which is defined as $$\phi(z)\triangleq\sum_{n=-\infty}^{+\infty}\gamma(\tau)z^{-n}$$ where $\gamma(\tau)$ is the covariance function. The author points out directly that the convergence range is a strip $r_1<|z|<r_2$ where $0<r_1<1$ and $r_2>1$. The spectrum function is apparently a Laurent series. I use d'Alembert's ratio test and find that the author's claim indicates that $$|\lim_{n=+\infty}\frac{\gamma(n+1)}{\gamma(n)}|<1 \qquad (1)$$ If the stationary process is Markovian Gaussian stationary process, then it decays exponentially. Other than that, is there a general proof of the convergence range of the spectrum of a stationary process? In other words, why does Eq. (1) always hold for stationary process. Thanks in advance for your help!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.