Repozytorium publikacji - Politechnika Gdańska

The problem of estimating spectral density of a nonstationary process satisfying local stationarity conditions is considered. The proposed solution is a two step procedure based on local autoregressive (AR) modeling. In the first step Bayesian-like averaging of AR models, differing in order, is performed. The main contribution of the paper is development of a new final-prediction-error-like statistic, which can be used to select optimal estimation bandwidth in the second step of the procedure. Simulation experiments demonstrate that the combined cooperative-competitive approach outperforms the previously introduced fully competitive scheme.

Autorzy

• dr hab. inż. Michał Meller link otwiera się w nowej karcie ,
• prof. dr hab. inż. Maciej Niedźwiecki link otwiera się w nowej karcie ,
• Damian Chojnacki link otwiera się w nowej karcie ,
• Adam Lasota link otwiera się w nowej karcie