Near second-order Markov property of MA-filtered first-order Markov processes

Consider a first-order Markov process observed at the output of a length-NN moving-average (MA) filter, with output covariance

C1MP+MA(k;α)=j=N+1N1(Nj)σ2eαkj.C_{\mathrm{1MP}+\mathrm{MA}}(k;\alpha)=\sum_{j=-N+1}^{N-1}(N-|j|)\sigma^2\,\operatorname{e}^{-\alpha|k-j|}.

Near second-order Markov conjecture. The MA filtering of the first-order Markov process generates nearly a second-order Markov process.

This concerns the effective Markov order induced by moving-average filtering of a process with exponentially decaying correlations. The statement is presented as a conjecture, but the supplied text does not specify what “nearly” means or provide evidence resolving it.

Sources & referencesView supporting material

Primary source

Pavel Loskot, “Polynomial Representations of High-Dimensional Observations of Random Processes”, arXiv:2010.07991 (2020).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.