Extrapolation Problem for Multidimensional Stationary Sequences with Missing Observations

Oleksandr Masyutka, Mikhail Moklyachuk , \corrauth, Maria Sidei · 2025

Read Extrapolation Problem for Multidimensional Stationary Sequences with Missing Observations by Oleksandr Masyutka, Mikhail Moklyachuk , \corrauth, Maria Sidei (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-07228.txt.

Introduction

Oleksandr Masyutka, Mikhail Moklyachuk, and Maria Sidei tackle extrapolation for multidimensional stationary sequences when some observations are missing. The goal is to estimate linear functionals that depend on unknown values of the sequence, using observations corrupted by an additive stationary noise sequence. Missing data turns a clean prediction problem into one where the estimator must be built around the gaps.

Under spectral certainty, where the spectral densities of signal and noise are known exactly, the authors derive formulas for the mean-square errors and the spectral characteristics of the optimal linear estimates. Under spectral uncertainty, with densities known only to lie in allowed sets, they apply the minimax robust method. They identify the least favorable densities and the minimax characteristics of the optimal estimates for special classes of allowed sets.

Catalog record

CREATOROleksandr Masyutka, Mikhail Moklyachuk , \corrauth, Maria Sidei
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE69 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-07228

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Extrapolation Problem for Multidimensional Stationary Sequences with Missing Observations Oleksandr Masyutka, Mikhail Moklyachuk , \corrauth, Maria Sidei Abstract This paper focuses on the problem of the mean square optimal estimation of linear functionals which depend on the unknown values of a multidimensional stationary stochastic sequence. Estimates are based on observations of the sequence with an additive stationary noise sequence. The aim of the paper is to develop methods of finding the optimal estimates of the functionals in the case of missing observations. The problem is investigated in the case of spectral certainty where the spectral densities of the sequences are exactly known. Formulas for calculating the mean-square errors and the spectral characteristics of the optimal linear estimates of functionals are derived under the condition of spectral certainty. The minimax (robust) method of estimation is applied in the case of spectral uncertainty, where spectral densities of the sequences are not known exactly while sets of admissible spectral densities are given. Formulas that determine the least favorable spectral densities and the minimax spectral characteristics of the optimal estimates of functionals are proposed for some special sets of admissible densities. \runningheads Extrapolation Problem for Sequences with Missing Observations O. Masyutka, M. Moklyachuk and M. Sidei \corraddr Mikhail Moklyachuk (Email: [email protected]). Department of Probability Theory, Statistics and Actuarial Mathematics, Taras Shevchenko National University of Kyiv, Volodymyrska 64 Str., Kyiv 01601, Ukraine. keywords Stationary sequence, mean square error, minimax-robust estimate, least favorable spectral density, minimax spectral characteristic AMS 2010 subject classifications. Primary: 60G10, 60G25, 60G35, Secondary: 62M20, 93E10, 93E11 DOI: 10.19139/soic.v7i1.527 1 Introduction The problem of estimation of the unknown values of stochastic processes is of constant interest in the theory and applications of stochastic processes. The formulation of the estimation problems (interpolation, extrapolation and filtering) for stationary stochastic sequences with known spectral densities and reducing these problems to the corresponding problems of the theory of functions belongs to Kolmogorov [ 17 ] . Effective methods of solution of the estimation problems for stationary stochastic sequences and processes were developed by Wiener [ 39 ] and Yaglom [ 40 , 41 ] . Further results are described in the books by Rozanov [ 36 ] , Hannan [ 12 ] , Box et. al [ 3 ] , Brockwell and Davis [ 4 ] . The crucial assumption of most of the methods developed for estimating of the unobserved values of stochastic processes is that the spectral densities of the involved stochastic processes are exactly known. In practice, however, complete information on the spectral densities is impossible in most cases. In this situation one finds parametric or nonparametric estimates of the unknown spectral densities and then apply one of the traditional estimation methods provided that the selected spectral densities are true. This procedure can result in significant increasing of the value of the error of estimate as Vastola and Poor [ 38 ] have demonstrated with the help of some examples. To avoid this effect one can search estimates which are optimal for all densities from a certain given class of admissible spectral densities. These estimates are called minimax since they minimize the maximum value of the error of estimates. The paper by Grenander [ 11 ] was the first one where this approach to extrapolation problem for stationary processes was proposed. Several models of spectral uncertainty and minimax-robust methods of data processing can be found in the survey paper by Kassam and Poor [ 16 ] . Franke [ 7 , 8 ] , Franke and Poor [ 9 ] investigated the minimax extrapolation and filtering problems for stationary sequences with the help of convex optimization methods. This approach makes it possible to find equations that determine the least favorable spectral densities for some classes of admissible densities. In the papers by Moklyachuk [ 22 , 23 ] results of investigation of the extrapolation, interpolation and filtering problems for functionals which depend on the unknown values of stationary processes and sequences are described. The problem of estimation of functionals which depend on the unknown values of multivariate stationary stochastic processes is the aim of the papers by Moklyachuk and Masyutka [ 25 ] - [ 27 ] . In the book by Moklyachuk and Golichenko [ 24 ] results of investigation of the interpolation, extrapolation and filtering problems for periodically correlated stochastic sequences are proposed. In their papers Luz and Moklyachuk [ 18 ] - [ 20 ] deal with the problems of estimation of functionals which depend on the unknown values of stochastic sequences with stationary increments. Prediction problem for stationary sequences with missing observations is investigated in papers by Bondon [ 1 , 2 ] , Cheng, Miamee and Pourahmadi [ 5 ] , Cheng and Pourahmadi [ 6 ] , Kasahara, Pourahmadi and Inoue [ 15 ] , Pourahmadi, Inoue and Kasahara [ 33 ] , Pelagatti [ 32 ] . In papers by Moklyachuk and Sidei [ 28 ] - [ 31 ] an approach is developed to investigation of the interpolation, extrapolation and filtering problems for stationary stochastic sequences with missing observations. In this paper we present results of investigation of the problem of the mean-square optimal estimation of the linear functional A\vec{\xi}=\sum\limits_{j=0}^{\infty}\vec{a}(j)^{\top}\vec{\xi}(j) which depends on the unknown values of a multivariate stationary stochastic sequence \{\vec{\xi}(j),j\in\mathbb{Z}\} . Estimates are based on observations of the sequence with an additive stationary stochastic noise sequence \{\vec{\xi}(j)+\vec{\eta}(j)\} at points j\in\mathbb{Z}_{-}\backslash S=\{\ldots,-2,-1\}\backslash S , where S=\bigcup\limits_{l=1}^{s}\{-M_{l}-N_{l},-M_{l}-N_{l}+1,\ldots,-M_{l}\} . The problem is investigated in the case of spectral certainty, where the spectral densities of the signal and the noise sequences \{\vec{\xi}(j),j\in\mathbb{Z}\} and \{\vec{\eta}(j),j\in\mathbb{Z}\} are exactly known, and in the case of spectral uncertainty, where the spectral densities of the sequences are not exactly known while a set of admissible spectral densities is given. We first propose results of investigation of the mean-square optimal linear estimate of the linear functional in the case of spectral certainty. To find the optimal solution of the estimation problem in this case we apply an approach based on the Hilbert space projection method proposed by Kolmogorov [ 17 ] and developed in the papers by Moklyachuk [ 22 , 23 ] , and Moklyachuk and Masytka [ 25 ] - [ 27 ] . We derive formulas for calculation the spectral characteristic and the mean-square error of the optimal estimate of the functional. Next, in the case of spectral uncertainty, where the full information on spectral densities is impossible, while it is known that spectral densities of the sequences belong to some specified classes of admissible densities, the minimax-robust method of estimation is applied. This method gives us a procedure of finding estimates which minimize the maximum values of the mean-square errors of the estimates for all spectral densities from a given class of admissible spectral densities. Formulas that determine the least favorable spectral densities and the minimax-robust spectral characteristics of the optimal estimates of the functional are proposed for some specific classes of admissible spectral densities. 2 Hilbert space projection method of extrapolation of stationary sequences with missing observations Let \vec{\xi}(j)=\left\{\xi_{k}(j)\right\}_{k=1}^{T},\,j\in\mathbb{Z} and \vec{\eta}(j)=\left\{\eta_{k}(j)\right\}_{k=1}^{T},\,j\in\mathbb{Z} , be multidimensional stationary stochastic sequences with zero mean values: E\vec{\xi}(j)=\vec{0} , E\vec{\eta}(j)=\vec{0} and correlation functions which admit the spectral decomposition (see Gikhman and Skorokhod [ 10

Read the complete text