The problem of weighted finite-rank series approximation of a time-series aimed at estimating the signal in the "signal plus noise" model, where the inverse covariance noise matrix is (2p + 1) diagonal, is considered in this paper. The problem of searching for weights that improve the estimation accuracy is solved. An efficient numerical method for the search for weights is constructed and theoretically validated. A numerical simulation is performed for several noise models to study the improvement of accuracy of the signal estimation method.

Translated title of the contributionПоиск весов для задачи оценивания сигнала конечного ранга в присутствии случайного шума
Original languageEnglish
Pages (from-to)236-244
Number of pages9
JournalVestnik St. Petersburg University: Mathematics
Volume54
Issue number3
DOIs
StatePublished - Jul 2021

    Research areas

  • Cadzow method, finite-rank series, Kullback–Leibler divergence, optimization methods, time series, weighted least square method, ALGORITHM, Kullback-Leibler divergence

    Scopus subject areas

  • Mathematics(all)

ID: 89719341