Partial Directed Coherence Asymptotics for VAR Processes of Infinite Order
Daniel Yasumasa Takahashi, Luiz Antonio Baccalá, Koichi Sameshima · 2008
Currently available asymptotic confidence interval/null hypothesis threshold results for partial directed coherence (PDC) are strictly valid for vector autoregressive (VAR) processes of finite order p. The present paper discusses the extension to more general situations when VAR(p) are used as approximations to more general VAR processes of infinite order, in which case the order of the best fitted model becomes a function of the number of available data points and the resulting confidence intervals must be rescaled appropriately.